Modeling Tree Diameter Growth Using Nonlinear Mixed-Effects Models

Abstract

A diameter-age model was developed for dahurian larch (Larix gmelinii. Rupr.) in northeastern China based on Chapman-Richards growth model using nonlinear mixed-effects modeling approach. The methods of model development include which parameters should be considered to be random and which should be purely fixed, as well as procedures for determining autoregressive correlation structures. Model performance was evaluated utilizing information criterion statistics (AIC, BIC, and LRT). The Chapman-Richards model with two random parameters showed the best performance. The inclusion of autoregressive models (AR(p), MA(q), ARMA(p,q)) in the mixed-effects model resulted in a significant improvement of model fitting statistics.