Abstract
Assessing potential site productivity is important for forest planning and predicting forest growth and yield. The most common measure of site productivity is the site index (SI), defined as the average height of dominant and co-dominant trees (top height) at a specified (index) age. Two types of SI curves were investigated in this study to build up a SI curve for a Sugi plantation in Zenlen area, Taiwan. Several models were used to compare the performance among models. Pairs of height-age observations were obtained through a stem analysis. Based on the mean square error criterion, the base-age-specific SI model proposed by Payandeh and Wang (1994) was ranked the best one with an average bias in percentage of -0.36. Despite the advantage of being invariant with the base age used, this study showed a large loss of accuracy and precision caused by base-ageinvariant models, especially for those trees younger than 10 yr. Moreover, the accuracy of the SI models varied considerably depending on the choice of predictor age for estimating the SI value at the base age. This study showed a tendency that the closer the predictor age was the base age, the higher the accuracy was that was obtained.