Classificação da capacidade produtiva de povoamentos de eucalipto a partir de parcelas temporárias

Abstract

The objective of this work was to perform the classification of forest sites from temporary plots. The database used to fit the models were composed by 12,050 records of age and average dominant height obtained from Eucalyptus urophylla and Eucalyptus grandis stands, located in Southern of Bahia, with an area of approximately 96,000 ha. For this, eight mathematical models were tested: Schumacher, Chapman-Richards, Weibull, Silva-Bailey, Prodan, Logistics, Clutter Jones and Mitscherlich. The models were evaluated by the standard error of estimate (Syx), graphic dispersion of residues through the histogram and normal probability of residual plot, Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). All models presented statistical values and graphics of residues with a close behavior to each other. Six site curves were constructed, with range of five meters, and the guide curve method using the reference age of seven years. Chapman Richards’ model was chosen, because it showed the best statistical results and is a biological model. The class III site had greater representation in the ranking per field, with a site index of 35.5 m.