A GENERALIZED MIXED-EFFECTS HEIGHT-DIAMETER MODEL FOR INTENSIVELY MANAGED Pinus taeda STANDS IN SOUTHERN BRAZIL

Abstract

Background: The insertion of stand variables allows for generalized mixed-effects height-diameter model, and the inclusion of random effects aids in model flexibility and application to new databases through calibration. This study aimed to develop a generalized mixed-effects model capable of accurately predicting the total height of Pinus taeda trees, subjected to different management regimes and at different ages.

Results: Experimental data between 4 and 27 years of age collected from stands of Pinus taeda located in the State of Santa Catarina were used. The experiment covers a combination of different planting densities and thinning intensities. The Lasso method was used to select the model’s independent variables, and mixed modeling was implemented, with calibration using BLUP (best linear unbiased predictor). Dominant height and quadratic mean diameter were selected in different formulations by the Lasso method to compose the predictor variables of the model, along with diameter at 1.3 m (dbh). Random effects were entered into the intercept and inverse of dbh terms of the model.

Conclusion: The results show that the generalized model developed presents flexibility and can be applied to Pinus taeda stands in southern Brazil using the calibration of random effects, when necessary, by collecting a small sample of plot data.