Height Growth and Site Index Curves for Inland Douglas-fir Based on Stem Analysis Data and Forest Habitat Type

Results of a stem analysis site index study for inland Douglas-fir ( Pseudotsuga menziesii [Mirb] Franco var. glauca ) growing in both even- and uneven-aged stands in the northern Rocky Mountains are reported. Separate height growth and site index equations were fit: a generalized logistic adequately described the height growth data, and a linear model arrived at by simplifying the results of Dahms' method adequately described the site index data. Forest habitat type proved to be a useful concomitant variable in all models, for curve shape differed between three habitat series groupings. The Richards function was not flexible enough to fit either the height growth or site index data, and Dahms' method produced models that contained twice as many parameters as needed. Furthermore, Dahms' method did not produce an acceptable height growth model. Weighted regression was used to fit all models. A first-order autoregressive variance component was also estimated, to account for autocorrelation; it was subsequently found that ignoring the autocorrelation problem did not result in biased parameter estimates. Precision curves illustrating the expected relationship between sample size, standard error, and age should help users achieve a specified error in estimating height growth and site index models. Forest Sci. 30:943-965.