Estruturação da produção de Tectona grandis Linn f. em Mato Grosso

• Ano de defesa: 2018
• Autor(a) principal: Favalessa, Cyro Matheus Cometti
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Recursos Florestais e Engenharia Florestal; UFSM; Programa de Pós-Graduação em Engenharia Florestal; Centro de Ciências Rurais
• Palavras-chave em Português: Modelos mistos; Sitio; Relação hipsométrica; Tabelas de produção; Mixed models; Site; Height-diameter relationship; Yield tables; CNPQ::CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL
• Link de acesso: http://repositorio.ufsm.br/handle/1/15729

The need to improve the management of forest resources, the growth modeling is fundamental to suport the decision-making regarding to the forest planning and yield structuring. The general goal int his research was to develop a diameters distribution model and yield structure of Tectona grandis in Mato Grosso. The research was developed in a forest stand of Tectona grandis in Nossa Senhora do Livramento, in the state of Mato Grosso. Regression models were tested to estimate the stand attributes such as the height-diameter relationship, the dominant height growth, the tree volume, the mean square diameter, the mean height, the number of trees per hectare and the minimum and maximum diameters. The reported probabilistic distribution models found in the literature and reestimative methods - which have already been valuated - and the parameters of the probability distribution model were also tested. Generic models were developed by stepwise and the insertion of variables with biological expression; the mixed model’s theory was also used to verify gains in precision and improvement in the behavior of the estimates when pertinent. The regression models adjusted by the ordinary least squares were evaluated according to the standard error of the estimate, adjusted coefficient of determination, inflation value of the variance, graphical analysis of the residues, as well as, the regression conditioners. The generalized linear models were evaluated according to the standard error of the estimation, the Baesyan and Akaike information criteria, the variance inflation value and the residuals graphical analysis. The probability distribution models were adjusted by maximum likelihood and selection of the best model was performed by grading the Kolmogorov-Smirnov, Chi-square, Anderson-Darling and Cramer-von-Mises grades. The reestimative parameters methods of parameter prediction, maximum likelihood regression and regression of accumulated density function were evaluated by the Kolmogorov-Smirnov test and the residues dispersion by each method. After selecting the best equations, a equations system was developed to estimate the diameters distribution and the yield struture through production tables. Height-diameter’s Stoffels equation with common angular coefficient and recalculated intercept for the central diameter tree was the best to express the behavior of the height-diameter relationship. The site curves presented monomorphic behavior by the Monomolecular equation. The Schumacher-Hall’s volume equation was the most adequate to estimate the trees volume. The equations developed to estimate the stand attributes were statistically consistent and expressed the biological behavior of the stand population. The Weibull function probabilistic distribution model was adequate to estimate the diameters distribution and the production tables developed can be used for the yield structuring of Tectona grandis stands.