Abstract
The analysis of the dominant height growth for Vochysia guatemalensis, based on a sample of permanent plots located in the Northern Huetar and the Caribbean Huetar regions of Costa Rica, allowed to distinguish 2 dissimilar growth patterns and develop differentiated site index curves. In order to carry out this work, all the information available about the growth of this species from public and private organizations was used, giving rise to a sample of 43 permanent sampling plots; which produced 397 dominant height-age averages, with which all models were adjusted. The first step was to classify the dominant height growth and as a result 3 site quality classes were preliminarily created; it allowed proving the existence of 2 different growth patterns in the sample: one represented by classes I and II (sample 1), and the other by class III (sample 2). Levene test did not reject the hypothesis of equality of variances between the samples, which indicates there is a high probability that Samples 1 and 2 may come from the same population, while t test rejected the null hypothesis of equality of means, revealing there are significant differences in the average of the dominant heights of both samples. Because of this, it was decided to adjust 2 separated sets of site index curves, based on the observed discrepancies. To build the curves, a base age of 9 years were used. In both cases, eight linear and non-linear (but linearizable) functions and 12 non-linear models were tested. For sample 1 the selected function was that of Bailey and Clutter, and for sample 2 the selected function was that of Prodan; both equations showed a good statistical fit and a good performance when graphed against the observed values. The equation for sample 1, presented the following statistics: r = -0.94; R2adj = 0.89; MSE = 0.0565; SEE = 0.2378; and the second equation for sample 2, showed the following results: r = 0.93; R2adj = 0.87; MSE = 0.3886; SEE = 0.6234.