Florística, fitossociologia e estimativa de variáveis florestais em um fragmento de cerrado stricto sensu , Gurupi-TO

Abstract

The Cerrado biome has a rich flora and presents a high level of endemism. In relation to the state context, the Tocantins is situated in the geographical transition zone between the Cerrado and the Amazon. The objective of this research was to evaluate floristic, phytosociology and application of sampling techniques for forest surveys, such as size and shape of plots, intensity and sampling processes in an area of Cerrado stricto senso. Through a census, in an area of 6.7 hectares, arboreal individuals with a limit of inclusion of chest circumference (CAP) greater than or equal to 15 cm were identified, as well as their total height measured. We sampled 15,434 individuals belonging to 78 species, 72 genera and 37 families of which the families Myrtaceae and Fabaceae predominated. The total density of the area and basal area were, respectively, 2,303 ind.ha-1 and 14,07 m².ha-1. In the study area the species that stood out in relation to phytosociological parameters were: Myrcia splendens, Astronium fraxinifolium, Magonia pubescens, Qualea parviflora, Protium heptaphyllum, Qualea multiflora, Vatairea macrocarpa, Tachigali aurea and Byrsonima stipulacea. After this step, variations of the sampling simulations were defined, considering the combinations: two sampling processes, random and systematic, different sizes of plots (250, 500, 1.000, 1.500 m²) and shapes (square, rectangular and circular) with the different sampling intensities (2,5, 5, 7,5 and 10%) evaluating three forest variables: volume, basal area and number of trees. In general, the increase in sample intensity showed a close relationship with the improvement of the estimates of the variables. However, it was verified that when using plots of 250 m², for the variables of interest, they presented tendencies of smaller values of the sampling and real errors, unlike what occurs in plots of larger sizes. Therefore, plots of 250 m² tend to more accurately capture the variation of the spatial distribution of the variables of interest, producing more reliable results than using plots of 1,000 and 1,500 m², under the same sampling intensity. So, using an intensity of more than 5%, with plots of 250 and 500 m², obtained a sampling error of less than 20%, an acceptable result for native forest area. Based on the above, it is recommended that, for the estimation of forest variables, small plots, 250 or 500 m² are used, and the sampling intensity should be as great as the available time and money resources allow