Comparação de modelos hipsométricos e descrição de um povoamento de Pinus taeda L.

Abstract

This article aimed to adjust a model that allows estimating the height, from measurements of the diameter, for a commercial plantation of Pinus taeda in Lages, Santa Catarina. The sample had 22 random circular plots of 400 m² each, and the individuals of P. taeda had their diameters and heights measured. Six hypsometric models were adjusted and, to select them, the adjusted R-Square (adjusted R²), standard error of the estimate (Syx), standard error of the estimate in percentage (Syx %) and graphical analysis of residual dispersion (relative error) were observed for each model. Also, the chi-square test was performed to validate the model that best estimated height. Descriptive statistics values were obtained for the diameter, height, dominant height, cross-sectional area and basal area, as well as the frequency distribution of the diameter. The Stoffels model was the best fit and was validated. Descriptive statistics indicated trees of high size, with homogeneity of data for height, diameter and dominant height. The diameter frequency distribution was close to normal, an expected pattern for an even forest.