Autodesbaste e diagrama de manejo da densidade em povoamentos de Pinus taeda L.

Abstract

This thesis aims at studying the efficiency of models which express the size-density relationship in even-aged stands of Pinus taeda L., handled in complete and full stock densities. For that, Reineke, Yoda, Zeide and Tang models were tested using data obtained from 50 permanent sample plots placed in stands implanted in spacing of 1.5x1.0 m, 2.5x1.0 m, 1.5x2.0 m, 2.0x2.0 m, 1.5x3.0 m, 2.5x2.0 m, 2.0x3.0 m, 2.5x3.0 m, measured every year until the age of 18 and maintained in full stocks. Results indicated that all of the size-density relationship models presented good statistical precision. However, the no-linear model of Tang was more efficient than the others, presenting excellent coefficient of determination (0.99), low standard error of estimate (0.0948), low coefficient of variation (1.17%), low bias (0.0086) and high efficiency (0,8976), in the estimate of the values of density of trees per hectare for average diameter, and the smallest considered value of the scores of the statistical criteria, equal to 6 points, with good distribution of residues of the number of trees per hectare. The self-tinning occurs in a dimension of average diameter directly proportional to the density of trees per hectare of planting. As bigger is the initial spacing, bigger is the average diameter in the moment of the beginning in the stand the self-thinning. The slope of the size-density and diameter equation varied per planting spacing, not allowing proving the universality of the self-tinning law, with value of this slope is equal -3/2. With help of the size-density model, 5 Indexes of Density of the Stand (IDS) were generated, with variation from the maximum 1400 to the minimum 600, with interval of class of IDS equal to 200, having as reference the standard diameter of 25 cm. The stand s Density Management Diagram (DMD), elaborated for the dendrometric variables average diameter, basal area and volume for hectare by IDS presented good efficiency with low mistake in the estimate of the values of these variables. Real volume and estimate volume per hectare in DMD presented an absolute difference of just -7.39 m3 ha-1 and a relative difference of -1.79 %, showing a great precision of the DMD model. The estimates volume of DMD per hectare, in relation to the real values, presented a value of efficiency equal to 0.99, what indicates a high precision, and a χ 2 value equal to 0.00034, not significant to 0.01% of probability, demonstrating that the real and the estimate values don't differ statistical amongst themselves.