Clustering of Amazon wood species based on physical and mechanical properties

• Autor(a) principal: Reis, Pamella Carolline Marques dos Reis
• Data de Publicação: 2019
• Outros Autores: Souza, Agostinho Lopes de, Reis, Leonardo Pequeno, Carvalho, Ana Márcia Macedo Ladeira, Mazzei, Lucas, Reis, Alisson Rodrigo Souza, Torres, Carlos Moreira Miquelino Eleto
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Ciência Florestal (Online)
• Texto Completo: https://periodicos.ufsm.br/cienciaflorestal/article/view/28114
• topic: Indication of species; Tropical wood; Wood technology; Indicação de espécies; Madeiras tropicais; Tecnologia da madeira

The intense search for consolidated native wood essences, which are highly used, can lead to the overexploitation of species and decrease their stocks in the forest. An alternative to this situation is the replacement of these species by others with similar wood properties and with sufficient forest growing stock. The objective of this work was to cluster the Amazon species through wood physical-mechanical properties and perform the discriminant analysis to identify which technological characteristics are more important for the clustering. The species studied came from eight different locations in the Amazon region. The properties used were: basic density, contraction (tangential, radial and volumetric), static flexion, compression parallel and perpendicular to the fibers, Janka hardness parallel and transversal, traction perpendicular to the fibers, cracking and shearing, all obtained from the national specialized literature. Multivariate Cluster analysis (simple Euclidean distance and Ward’s method) and the discriminant analyses were used to evaluate the clustering. Cluster analysis was efficient to cluster the species, which were separated into three distinct groups. The species that stood out were Helicostylis pedunculata Benoist. and Tachigali chrysophylla (Poepp.) Zarucchi & Herend., f or clustering with the most commercialized species. The lowest values of Wilks’Lambda were wood density (0.759053), shearing (0.802960) and compression parallel to the fibers (0.825594). These characteristics were the most determinant to discriminate the clusters. The clustering analysis was efficient for the separation of the species into marketing clusters