Alternativas de modelagem para a estimativa da altura de eucalipto Modeling alternatives for height estimate of eucalypt

Abstract

This work aimed to evaluate alternatives for generalist hypsometric modeling of eucalypt via linear regression, with inclusion of distance-independent competition indices as a predictor variable. The data used came from 34 plots distributed in four 72-month-old forest management units. Seven distance-independent competition indices were calculated. The predictive performance of 16 generalist hypsometric models was evaluated, and 14 double-entry models (DAP and competition index) were proposed. All equations generated were biologically consistent. Given the information absence on the height of dominant trees, generalist modeling can be better employed with inclusion of the IC4 index, which represents basal area of neighboring competing trees. It is concluded

that the inclusion of competition indices in hypsometric models increase generalization capacity of equations for eucalyptus stands with different productive capacities. The IC4 competition index favors fit quality and predictive performance of the hypsometric modeling in the studied sites.

Keywords: Competition, generalization, hypsometric relationship, linear regression.