Modelagem para prognose precoce do volume por classe diamétrica para Eucalyptus grandis

Abstract

The present study was designed to model the features of forest: site, survival, generic hypsometric ratio, minimum, maximum diameter, variance of diameters, basal area and arithmetic mean of diameters, to adjust and select probabilistic functions as Weibull, Sb and Beta by different adjustment methods and evaluate the veracity of the prognoses and the model potency to accomplish them early. The study was developed on data of Eucalyptus grandis from the Empresa Celulose e Papel Votorantin - CELPAV- Florestal S/A , situated in the town of Luiz Antonio, in the state of São Paulo. The data of the present work were collected from 5 projects of the enterprise, comprehending the regions of Guatapará (region 1), Luiz Antonio (region2), Altínopolis (region 3), Tamoio (region 4), Uberlandia (region 5). The development of the study was composed of five phases. In phase 1 was done the site classification by utilizing the method of the algebraic difference, evaluating the performance of the models in their anamorphic and polymorphic form. Phase 2 was the development and selection of the models of the features of the stand, being that the selection criterion of the most exact equation was through the determination coefficient (R2) , residual standard error (Syx) and graphical distribution of residues. In phase 3, the hypsometric models and fractional and integer potency polynomial were utilized for estimation of the volume. Phase 4 was the adjustment and selection of the diameter distributions (Beta, Weibull and Sb Johnson, in the several adjustment forms); and form 5 was to obtain the prognosis of growth and production of volume. The main results found were: the equations selected for the features of the stand presented themselves in an effective form as the statistics obtained were satisfactory, providing exact estimations; the distribution function Weibull adjusted by the method of the moments, proved to be superior to the other methods and the optimum value of “a’’ was of 5% of the minimum diameter; the early prognosis of the volume for region 1 (Guatapará) for the site index 25,5m should be done from the age of 24 to 48 months for other ages. When the prognosis ages were close to the rotation, the latter should be become viable by means of the volume equation adjusted by any one of the criteria adopted . But for the 28,5m site, the prognosis may be accomplished from the 26 to 33 month class for any other age. When the prognosis ages be close to the rotation , it must become viable through the tapering equations obtained from the cubage per age class or per cubed trees at different ages. Also, from the age classes 34 to 48 months and 9 to 58 months in the two years evaluated, it is perfectly possible to perform prognosis for other ages. But for region 2 (Luiz Antonio), the early prognosis of the volume for site index of 25,5m may become viable from the age class of 26-33 months for any other age. Nevertheless, when the prognosis is accomplished for the ages close to rotation the thinning function should be utilized, generated from trees rigorously cubed at different ages. For the site index of 28,5 m the early prognosis is possible from the age classes of 34 to 48 months for the other ages. When the prognosis age be close to the rotation, the latter should become viable through the stand equations (VCP and present). The early prognosis, proved efficient in estimating the volume , providing average errors of 2,39%and 4,82% in the estimates of the volume in the site indices of 25,5 m and 28,5 m, respectively.