Avaliação de modelos polinomiais segmentados e não segmentados na estimativa de diâmetros e volumes comerciais de Pinus taeda

Abstract

This study objectived to evaluate the accuracy of three segmented polynomial taper models for estimating diameters and volumes along the stem of Pinus taeda; and to verify if the more accurated model for diameters estimates is also the more accurated for estimating partial and total volumes. The database used was composed of 58 trees of Pinus taeda cubed by the Smalian’s method, in the farms of “PISA – Papel de Imprensa S/A” enterprise, located in Jaguariaíva county (PR). The 58 trees were separated in 8 diameter classes. The segmented models proposed by Clark, Souter and Schlaegel (1991); Max and Burkhart (1976); and Parresol, Hotvedt and Cao (1987), were fitted for each diameter class and for the total group of data. The accuracy of the models in estimating diameters and volumes along the stem of Pinus taeda was evaluated through the following statistics: coefficient of determination; residual standard error; graphic analysis of the medium profile of the stem; mean deviation of the diameters and volumes for each position along the stem; standard deviation of the differences of the diameters and volumes; sum of square of the relative residue; and percentage of the residues of the diameters and volumes. Starting from the last four statistics, it was elaborated a ranking to detect the more accurated model in each position of measurement along the stem. It can be pointed out that the model of Clark, Souter and Schlaegel (1991) is indicated for estimating diametes along the whole stem profile, and total and parcial volumes of Pinus taeda; the model proposed by Max and Burkhart (1976) is the second best option for estimating total and partial volumes, even though its estimates are not as accurated as the estimates improved by Clark’s equation. The model of Parresol, Hotvedt and Cao (1976) doesn’t presented accurated estimates of diameters and volumes along the stem in any diameter class.