Height-diameter relationships in sweetgum (Liquidambar styraciflua)- dominated stands

This report presents height-diameter equations for five hardwood species commonly associated with sweetgum (Liquidambar styraciflua L.) stands within the Piedmont and Coastal Plain regions of Alabama, Georgia, and South Carolina. Mixed-effects regression and the Chapman-Richards function were used to fit height-diameter equations for each species. Two equations were developed for each species, a base model with diameter as a predictor and a covariate model that included diameter and measures of stand structure. Compared with the base model, a marginal improvement in fit was seen in covariate models that include stand basal area and/or basal area of trees larger than a subject tree. Models were evaluated using an independent data set, and standwise calibration was used to estimate random parameters for the evaluation stands. Calibration improved prediction of tree height for all species, and increasing the number of sample trees used in the calibration beyond three had a minimal effect on improving prediction. This study supports previous findings that height prediction improves as one moves from a diameter-based model to a locally calibrated covariate model.