Modeling Individual-Tree Height Prediction for Semi-Evergreen Forest: MultilevelLinear Mixed Effect Model Approach

Individual-tree height models were developed for dipterocarpaceae and nondiptoracarpaceaetree family for semi-evergreen forest in Seam Reap, Cambodia. Tree variables data were collected from long term permanent sample plots for all species. The models were first fitted using multiple linear regressions since it is the most commonly used statistical method in forest modeling. The models then were fitted using multilevel linear mixed-effects model due to correlated measurements of tree height over time. Both models were compared using validation data with two statistics calculations; prediction error and prediction bias. The results indicated that the mixed-effects model performed better than the regression model.