| Summary: | The objective was to adjust and select a mathematical model to predict the stem volume of Tectona grandis L. f. in an 11-year-old plantation in Nuevo Urecho, Michoacan. The sample analyzed was 128 trees, in which it was measured the stump diameter, normal diameter and diameters to different sections, from stump height to total height. Six stem volume models were adjusted using the SAS 9.2® statistical package Model procedure. The selection of the model was made from the total square error (SCE), the root mean square error (REMC) and R2adj as well as the significance of their parameters. Compliance with the assumptions of normality, homogeneity of variances and independence of waste frequency was verified. A rating criterion was generated that ranked each statistics according to the order of importance, with the best being those with the lowest value in the overall rating. Meyer's model was the best fit with SCE= 0.0399, REMC= 0.0179 and R2adj= 0.8246. The Shapiro-Wilk test (0.97), White´s Test (13.00), Durbin-Watson (2.0112) do not violate the regression assumptions and guarantee the effectiveness of the model for the estimation of the dependent variable. This model is considered appropriate to predict more accurately the stem volume as a function of the normal diameter (d) and total height (AT) of T. grandis in plantations with physical and biological conditions similar to those of the site of this study.
|
|---|
