Form factors vs. regression models in volume estimation of Pinus taeda L. stem
Volume is an important information to know the available potential of a forest stand, although direct volume determination is expensive and often impractical. Thus, precise and accurate indirect methods are fundamental. Three volume estimation methods of Pinus taeda stem were compared: mean form factor, form factor by diameter class, and volume models fitted by regression. A total of 146 trees were used for this study, in which 96 were used for fitting and estimation, and 50 for validation, collected in the middle of the Paraná State, Brazil. Fitting statistics (R2aj and Syx%) were used to evaluate regression models. Bias, precision and accuracy statistics, t-test and residue graphical analysis were used to judge estimation methods. Estimation by mean form factor (diameter class) and the Spurr model were equivalent, being superior to the others examined. It was concluded that the use of mean form factor by diameter class has satisfactory statistical performance, robustness and simplicity in volume estimation.
Os autores permanecem com os direitos autorais de tudo que publicarem na Científica. Opiniões e conceitos contidos no artigo e a fidedignidade e exatidão das informações e das referências nele apresentadas são de exclusiva responsabilidade dos autores. A reprodução parcial ou total dos trabalhos é permitida desde que seja explicitada a fonte de referência.