Modelagem volumétrica em povoamentos de Eucalipto

Abstract

The volume and tapering of tree stems, associated to wood quality, are the most relevant variables to determine the value and use of the product. The evaluation and development of models are essential for the adequate prediction of these variables. This work was developed with the intent of modeling the volume of Eucalyptus spp. at the tree level, using the variable-exponent functions, and at the stand level, using the growth and forest yield models. In the first article, two alternatives were evaluated to estimate the growth and yield of Eucalyptus stands based on modifications of the original Clutter’s model and in the data structure. Alternative 1 consisted of testing distinct transformations of the proposals in the original Clutter model for the site and basal area, and adjust the complete modified model (double and triple interactions). For this alternative, all sample plots were used, and the longitudinal trait of the database was not considered. The equations for basal area and volume resulting from alternative 1 were transformed into system and adjusted for the data of the sample plots with two or more measurements (alternative 2). These alternatives were compared to the system originally proposed by Clutter. The Clutter system was the best alternative. The system proposed in alternative 2 presented statistics similar to those of the Clutter system. The basal area equation of alternative 1 presented low precision. However, the volume equation presented precision similar to that of the Clutter system. The prediction errors for volume and basal area for the validation data were higher in alternative 1. In the second article, the quality of the Kozak variable-exponent equation adjustments were evaluated in order to estimate the diameters along the stem and total volume (outside and inside bark) of Eucalyptus spp. trees. A new proposal to estimate the relative height of the inflection point – p by means of geometric methods was presented to avoid subjectivity during the determination of p. The variable-exponent functions were adjusted, considering different strategies for the determination of p (as parameter, according to that defined by Kozak and mean value estimated by geometric methods), and a new exponent variable-equation was developed. The adjusted equations presented good precision and good ability for predicting diameters and volume. The strategy used for determining p had no influence over the precision of the equations. The new proposed equation presented behavior similar to those already developed by Kozak, however, with the advantage of presenting a simpler exponent.