ABSTRACT
The objective of this study was to apply and evaluate methods to estimate multiple volumes in a Community Management Area, located in the Tapajós National Forest. 516 sample-trees were rigorously scaled of 21 commercial species, all with 50 cm > dbh < 160 cm. The unique model for multiple volumes (regression analysis) proposed by Leite et al. (1995) and Artificial Neural Networks (ANN) were used to estimate the volume of stem and branches. After preliminary results using the regression method, it was decided to stratify the data into three parts, considering the relationship between volume of branches and volume of commercial stem. The equations generated showed improvement in the adjustments. However, the results with ANN were better than the ones generated by regression equations.
Keywords:
Neural networks; Multiple use; Forest residues
