Modelagem volumétrica para colmos de Dendrocalamus giganteus WALL. ex Munro (Poaceae)

Scientific investigations on the volumetric quantification of bamboo are essential in defining sustainable management strategies. This work aimed to evaluate different regression models for volume estimate of Dendrocalamus giganteus Munro (Poaceae) culms. Rigorous cubing of 16 culms was realized within the intervals of 6.05 to 17.28 cm dbh and 10.3 to 28.0 m total height. Fifteen regression models were tested to woody and apparent volumes estimate. Equations obtained from the Spurr and Schumacher-Hall logarithmic linear models presented all significant parameters using t test (p ≤ 0.05) and were more accurate, with few deviations and strong correlation coefficient (𝑟 𝑌𝑌^> 0.96; p ≤ 0.05). It is concluded that logarithmic transformations of variables is an operationally viable tool in the search meeting of ordinary least squares method premises and predictive improvement of linear models. Nonlinear models linearization through application of logarithmic properties is a promising alternative for expressing statistical effect of specified parameters. Adjustments of Spurr and Schumacher-Hall models can be used in their logarithmized form, accurately and without bias.