Bark volume and weight modeling of Mimosa scabrella Bentham trees in the Curitiba metropolitan region

Abstract

The objective of this paper was to model bark volume and weight of Mimosa scabrella Bentham trees. To achieve this objective 440 trees were cubed by the Hohenadl’s method using 10 sections. These trees had ages ranging from 6 to 17 years, and were located in several counties inside the Curitiba (Parana) metropolitan region. The stems from 194 trees were weighed in the field. Samples of bark from these trees were collected, weighed in the field (green weight) and brought to the laboratory for drying until constant weight (dry weight); thus it was calculated the relationship dry weight/green weight of the sample and extrapolation for the whole stem. For the modeling, several mathematical models from the literature for estimating dry weight of Acacia mearnsii were selected. New models based on the forward’s method were also built. Twenty different models were adjusted using the 1/d²h to weight the arithmetic models with two entries and 1/d² for the models with single entry. The best models for each case were selected basing on the adjusted coefficient of determination (R²aj), standard error of estimate in percentage (syx %) and on the residual graphic distribution. In general, the new models built with variables most correlated with the respective independent variables were the most precise. The best fitting were obtained for bark volume and green weight, with syx% ranging from 15 to 18%, and R²aj from 0.96 to 0.98. The best equation for estimating bark volume in percentage presented R²aj = 0.47 and syx% = 17%.The best models for estimating bark dry weight presented higher syx% and lower R²aj than those obtained for green weight.