Uso do modelo linear multivariado de crescimento no estudo da densidade básica da madeira de eucaliptos

• Ano de defesa: 2014
• Autor(a) principal: Batistela, Gislaine Cristina [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Analise multivariada; Eucalipto; Madeira - Densidade; Wood
• Link de acesso: http://hdl.handle.net/11449/110956

In biological situations in which there is structural dependence within experimental unit and observing a more accurate statistical viewpoint it should be accounted not only data variance but also the covariances (variations among characteristics within the same unit). In this study it is presented analytical procedures data for such situation using MANOVA technology (Multivariate Variance Analysis), growth multivariate linear models (MLMC) for wood basic density estimation from eucalyptus considering 5 sample heights on tree trunks from base to top (sampled disks 0% - base -, 25%, 50% 75%, 100% of tree commercial height) and also comparing the precision of polynomial regression model estimators to those obtained from MLMC technique. Used data set have basic density values from three groups of eucalyptus trees being Eucalyptus saligna, E. grandis e E. grandis x E.urophylla, with =27, =31, =30 experimental units, respectively. Results showed that there is low usage of MLMC within Agronomic and Forestry Sciences; that the tree top commercial position (100%) is the differentiating for all in the groups being the bottom position inappropriate for differentiation and that the general data variability indicates the necessity of models which should consider the structural dependence in the study of wood basic density.