Hypsometric approaches to Eucalyptus experiments in Brazil

Abstract

The objective of this study was to determine the best method to estimate tree height by different hypsometric approaches. This study is a provenance trial with Eucalyptus spp. in a 35-year-old pioneering plantation in the municipality of Lavras, Brazil. The census was taken by measuring the diameter at 1.30 m above ground and total height. All x and y coordinates of the individual trees were obtained and these trees were classified according to their stem quality. The ordinary kriging, cokriging and ordinary kriging of the residuals generated from the regression models were fitted by an exponential model. The Curtis model was selected for calculating the regression. The best method to estimate the height-diameter relationship was based on the statistical and graphical criteria. The spatial prediction model did not adequately represent the height-diameter relationship. The regression technique was more precise and accurate by inserting variables that helped capture different development standards of the trees. The regression-kriging produced more accurate total height estimates. Inclusion of the stochastic effect in the general spatial behavior of the total height variable helped to capture specific details about the stand. Therefore, this method is recommended for experimental areas.