ABSTRACT
In this paper, six Populus deltoides Marsh. clonal forests planted in 1986 were structurally evaluated between 1987 and 1995 (1 to 9 years old). Probabilistic functions normal, gamma, beta, Johnson SB and Weibull with two and three parameters were fitted to the observed diametric distributions data, by means of, depending on the case, the Moments Method, the Maximum Likelihood Method, the methodology proposed by Loetsch et al. (1973) and the mean and standard deviations to estimate the corresponding parameters. Goodness of fit was evaluated considering the relative deviations between the observed and predicted diameters sum raised to the first, second, third and fourth power. For all analyzed functions, deviations increased with the power increment, and the Weibull distribution with two parameters (a = 0) presented the smaller values for all the observed years considering diameters raised to the fourth power. Values of relative deviations between observed and predicted diameter sums raised to the fourth power were lesser than 0.005 (0.5%) for the Weibull function with two parameters, excepting the first year, in which all functions presented higher deviation values (between 11% and 13%). Weibull scale (b) and shape (c) parameters were estimated through linear regression fitted equations using the stand quadratic mean diameter as independent variable.
Key words:
weibull function; simulation; forest management