Resumo:
The data came from stem analysis of 92 dominant trees, and from the measurement of dominant height in 324 temporary sample plots from the National Forestry Inventory, distributed all over the States of Paraná and Santa Catarina. The oldest trees came from the National Forest of Irati. The sample covered all sites and ages from 1 to 27 years old. Seven regression models were tested to fit the dominant height - age relationship. Statistical and residual analysis indicated that the Chapman - Richards biological model was also the best for Pinus taeda. A linear regression with intercept zero was used to fit site index values over dominant height for several ages. The linearity was found and indicated the validity of the derived site index curves. The estability of the derived site index curves was also tested by using a distinct set of data comming from 43 stem analysed trees. All applied tests gave evidences that the developed set of site index curves will provide unbiased estimates of height growth of dominant trees for all site classes. Therefore they can be used for site classification for Pinus taeda in the States of Paraná and Santa Catarina, and very likely for the States of Rio Grande do Sul and São Paulo where this species is also planted. Meanwhile tests should be done for using these curves for another regions outside the places where the sample was taken.
