A Preliminary Analysis of Diameter Growth Based on the Long Term Measurements on Permanent Plots

Abstract

The availability of Poisson regression model and linear mixed model for diameter growth modeling was examined comparing with ordinal regression model. Periodic diameter growth was modeled using initial diameter as the explanatory variable by Poisson regression model. The result showed that Poisson regression model was not superior to ordinal regression model. Also, linear mixed model was used to explain the variance of diameter growth. Using linear mixed model, the goodness-of-fit of diameter growth model was improved and it was suggested that coefficients of regression were estimated more precisely. This study suggested that linear mixed model should be used to analyze diameter growth derived from repeated measurements of permanent plots.