Hypsometric relationship modeling of a mixed eucalyptus plantation in Santa Catarina, Brazil
DOI:
https://doi.org/10.35977/0104-1096.cct2022.v39.27149Keywords:
adjustment, dendrometry, statistics, forest management, regressionAbstract
The hypsometric models allow of the estimation of the total height of trees through the relationship of this variable with the diameter at 1.30 m from the ground (diameter at breast height – DBH). The present work aimed to evaluate the quality of fit of 18 linear models of the hypsometric relationship for a mixed stand of Eucalyptus spp. planted in Canoinhas, SC. The following selection criteria were applied: the graphic analysis residuals; the estimation error (Syx%); and the determination coefficient adjusted by the Schlaegel´s index (R²aj). The multicollinearity between the independent variables was a problem for some models with more than one independent variable. All models showed similar height estimates for the intermediate DBH values; however, some of these models diverged as to the larger diameter classes. The Henriksen, Näslund, Petterson, Sanquetta, power, Schumacher, and Wykoff models were the most suitable ones to estimate the total height as a function of DBH because they combined the fewest estimation errors with the highest adjusted coefficients of determination; from these seven models, it was not possible to select even one because of the similarity of the values of the selection criteria. Despite the weak relationship between the studied variables, the estimated heights did not differ statistically from the measured heights.