Geoestatística aplicada ao manejo florestal experimental em floresta ombrófila mista

Abstract

Forests present spatial-temporal strucutre, and their management can be aided by geostatistics. The present study aimed to use geostatistics in the experimental forest management of Mixed Ombrophylous Forest (MOF), in Rio Grande do Sul, Brazil, with two case studies. The specific objectives were to determine areas of production for a population of Araucaria angustifolia and check the sensitivity of geostatistics to different intensities of management (selective wood harvesting), at different time points, before and after the intervention in the forest. The first study was carried out in an area of 11.35 ha in Tapera, using census data from a population of Araucaria, which was used as a virtual sampling. Punctual ordinary kriging and co-kriging were used to the data of 52 virtual sampling units (30x30m) obtained. Cross semivariograms were adjusted based on the spatial structure of the number of individuals for basal area (G), volume (V), biomass (B) and carbon (C) combined through the use of map algebra to determine the production zones (PZ). The second study was held in Tupi Farm, Nova Prata, using sample units of 0.50 ha, with subunits of 10x10 m, where selective wood harvestings were implemented in 2002, with the removal of 0 (control), 20 (light harvest), 40 (medium harvest) and 60% (heavy harvest) of basal area in all diameter class. Inventories were carried out in 2001 (pre-harvesting), 2006 and 2010 (1st and 2nd monitoring). The available data were basal area and commercial volume, organized by subunits. In the first study, low, medium and high production zones were obtained (55.03, 35.54 and 9.43 % for the area of forest fragment, respectively). We obsereved that the forest was under disturbance and the population had balanced diameter distribution. In the second study, the light harvesting caused the less changes in the spatial structure of the forest, more noticeable in the simulated surface relative to the semivariogram, with the replacement of the wood removed when compared to the others. The control area was not more structured than the light harvesting, besides producing less wood. To the medium harvesting we observed pure nugget effect because it intensified the existing randomness in the sample unit prior to the intervention. However, in the heavy harvesting, there were major changes in the forest structure, where areas of high basal and commercial volume areas have become low value areas due to the mortality of individuals remaining in the former, and to the increase and inflow of trees occurring in the latter. The light selective harvesting was the most suitable, and it was spatially less structured, but more productive when compared to the control. Therefore, geostatistics may be used in forest management since it detects changes in the spatial structure of the forest and describes the behavior of variables.