Estudo de funções de afilamento para representar o perfil e o volume do fuste de Pinus taeda L.

• Autor(a) principal: Ferreira, Maria Zélia
• Data de Publicação: 2004
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/11807
• topic: Modelos sigmoidais; Modelos polinomiais; Função de afilamento; Identidade de modelos; Sigmoid models; Taper function; Identity of models; Recursos Florestais e Engenharia Florestal

The accuracy of sigmoid and polynomial models for estimating diameters and volumes, along the trunk of Pinus taeda, in different production sites; for verifying if the best model to estimate diameters is also the best one for estimating volumes and for testing the identity of the models was studied in several production sites. The used database was composed by 278 trees of Pinus taeda rigorously scaled by the Smalian method, in areas belonging to InpacelIndústrias Arapoti S.A Company, in Arapoti-PR County. The sigmoid models tested were: Ormerod (1973) and its modification developed by Guimarães & Leite (1992), Kozak et al. (1969) and Demaerschalk (1972). The polynomials models tested were: Polynomial of the Fifth Degree, the Polynomial of Fractional and Whole Potencies of Hradeztky (1976) and the polynomial proposed by Goulding & Murray (1976). The accuracy of the models was evaluated by the following statistics: medium deviation in each measurement position along the trunk, standard deviation of the differences and sum of squares of the relative and percentile residues. All sigmoid models studied presented tendentious estimates of the diameters and of the volumes along the trunk of the trees, presenting only a few points of the required accuracy. The Polynomial of Fractional and Whole Potencies, in all production sites, was the best model for estimating diameters. The Polynomial of the Fifth Degree was the best model for estimating volumes. The best model for diameter estimates was not the best for volumes estimates. When estimating the identity of models, it was noticed that the better a model adjusts to the data; the more rigorous is the identity test.