Assessing alternatives for growth and yield modelling in eucalypt stands

Abstract

Growth and yield estimates are extremely important for forest stand management. The application of simple models is often useful, as it allows the use of plots with few or a single measurement. The objective of this study was to evaluate methods for modelling eucalypt stand volume under different conditions of climate and soil. We fitted the Gompertz, Logistic and Exponential models, stratifying by project and for the entire data set, to model the dominant eight and volume based on the usual validation statistics and residue analyses. To model growth and yield, five alternatives were evaluated: the Gompertz, Logistic and Exponential models, exponential-modified with the inclusion of the variable site index (SI) and the Clutter model. The best model was selected based on the validation statistics and the histogram of the relative error (RE). We conclude that the stratification by project presented greater accuracy of forecast. The best model to estimate yield was the exponential model modified with the inclusion of the variable SI. The technical age of cutting ranged from 36 months, in areas of higher productivity, to 84 months, in areas with low productivity.

Keywords: dominant height; forest management; measurement; site index