HYPSOMETRIC EQUATIONS FOR UNMANAGED Eucalyptus spp. IN OLD AGE WITH TECHNIQUES FOR THE INCLUSION OF COVARIATES
The aim of the study was to establish hypsometric equations for unmanaged Eucalyptus spp. in old age. For this purpose we measured the diameter and height of 513 stems distributed in 11 species and the hypsometric relationship was established by six regression models, being selected the one with the...
| Autores: | , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2016 |
| País: | Brasil |
| Institución: | Universidade Federal de Lavras (UFLA) |
| Repositorio: | Cerne (Online) |
| Idioma: | inglés |
| OAI Identifier: | oai:cerne.ufla.br:article/1095 |
| Acceso en línea: | https://cerne.ufla.br/site/index.php/CERNE/article/view/1095 |
| Access Level: | acceso abierto |
| Palabra clave: | Regression covariates decomposition of parameters. |
| Sumario: | The aim of the study was to establish hypsometric equations for unmanaged Eucalyptus spp. in old age. For this purpose we measured the diameter and height of 513 stems distributed in 11 species and the hypsometric relationship was established by six regression models, being selected the one with the best Akaike Information Criterion (AIC), standard error of estimative (Syx), Maximum Likelihood Ratio Test and Residual Graphical Analysis. Subsequently, the best model has undergone the inclusion of the covariates stem quality (Qf) and Species (Sp) by means of the decomposition of its parameters. Under these conditions, the model of Chapman and Richards showed the best performance in both modeling approaches. When compared both models, we observed a reduction of 71 AIC units and 7.4% in Syx and a significant improvement in all aspects of the residual distribution in the model with covariates. The results show that it is possible to provide hypsometric equations suitable for unmanaged Eucalyptus in old age, with and without addition of covariates, and the last technique has provided significant improvement in the quality of fit of the models. |
|---|