Influence of the procurement site on physical and mechanical properties of cupiúba wood species

Abstract

Studies that estimate technological properties of tropical wood species (especially those from the Amazon Rainforest) for their use in building construction, mainly structures, are very desirable. This paper aimed to investigate, aided by ABNT NBR 7190 (1997) recommendations, by Kruskal-Wallis analysis of variance (ANOVA), and regrouping bootstrap simulation technique, the influence of procurement sites (Caracaraí and Bonfim do Sul, State of Roraima, Brazil) and Cláudia (State of Mato Grosso, Brazil) on physical and mechanical properties of a Cupiúba wood species (Goupia glabra Aubl.). It was intended to assess the possibility of estimating (by linear, exponential, geometric, and logarithmic mathematical models) the physical and mechanical properties investigated as a function of density at 12% of moisture content. The results of ANOVA indicated equivalence in 94% of the properties of the Caracaraí and Claudia sites, and no equivalence in 50% of the properties in the Bonfim site; even after extrapolation by the bootstrap simulation technique, the non-equivalence was still 44%. Results obtained from the regression models implied a possibility of an estimate of the physical and mechanical properties of Cupiúba wood species using density as the estimator.