Avaliação de metodos de prognose da altura aplicados em Pinus elliottii da Floresta Nacional de Capão Bonito-SP

The present research was aimed at studying the height growth of Pinus elliottii from Capão Bonito, São Paulo State, due to the importance of knowing about this development while using a growth function for the prognosis of forest production. The objectives of the study were the following: a) to research different methods of prognosis through a bibliographical review and to select the ones that utilize growth functions which can be adjusted to a straight line, as these do not require special methods for the calculation of their coefficients; b) to obtain the height growth data from stem analysis of 30 year old trees; c) to estimate the values of height till the 30thyear of age simulating prognosis by adjusting the growth functions in different data periods or the following ages: 8, 10,12,14, 16, 18, 20, 22, 24, 26 and 28; d) to evaluate the results of the prognosis simulation by comparison with real data obtained from stem analysis. Prognosis of thirteen growth functions were tested and compared with the real growth up to the 30th year, being the age periods: 8, 10, 12, 14, 16, 18, 20, 22, 24, 26 and 28.The residue analysis test was used through associate probability, and its was confirmed that among the tested models only four had not tended to the probability value of 0,05 in all the prognosis periods. 6. H = H (I - 1) + e(a + b(I - 1) + c ln(I - 1)) 7. H = H (I - 1) + e (a + b ln(I -1) + c lnH((I - 1)) 10. H = H (I - 1) + e(a + b(I - 1) + c ln(I - 1)) 11. H = H (I - 1) + e(a + b(I - 1) + c ln(I - 1)) It was seen that the tendency of the curves presented by the functions of best adjustment began to get near the real data when the adjustment was made during the data period which comprises the young and the adult phasis, in other words, from the CAI (current annual increment) point of culmination. Through graphic analysis it was seen that, concerning prognosis, the tendency of the functions is not significantto the obtained results, since the tendentious functions also offered good prognosis at adjustment periods. Among the tested functions the one introduced by Schumacher (1939) [log II = a + b 1/I) was the only not to present a good adjustment at any of the studied prognosis phases, having been tendentious at all the adjustment periods except at the 8th year. The empyrical method of Backman was studied with the use of the probability graph paper and it has been concluded that this method it the safest for prognosis, since once the CAI point of culmination is related to the final theoretical heights, thus defining the growth tendency, it doesn't result in errors such as the ones wich occur at the use of growth functions while assuming the data tendency according to the adjusted mathematical functions