Description of height growth of hybrid eucalyptus clones in semi-arid region using non-linear models

Abstract

Brazil stands out worldwide for planting homogeneous forests, mainly pine and eucalyptus. Forestry production is of great importance for the country’s economy, being also a reference in sustainability, competitiveness and innovation. Of the 10 million hectares of planted trees, 76.3% is composed of the genus Eucalyptus, which makes Brazil one of the largest producers of this genus in the world. The analysis of the growth trajectory of trees of this genus can be a great ally in improving the management plans currently used. In this sense, the aim of this study was to compare the performance of the nonlinear models Gompertz, von Bertalanffy, Brody, Chapman-Richards and Schöngart, which were fit using the R software considering the first order autoregressive error structure (AR1), applied to data of average height, in meters, in relation to time, in months, totaling 15 observations obtained during six and a half years. Nonlinearity measures were used to check the adequacy of the linear approximations of models and as criteria for model selection the R2, AICC and DPR, with the Schöngart (AR1) model being the one that best fit the data.