Estratégias de amostragem em ecossistema sucessional de floresta madura nos planaltos interfluviais do alto rio Arapiuns, Amazônia Oriental.

Abstract

Tropical Forest ecosystems are responsible for the largest proportion of the world's forests, in which they need efficient techniques to quantify their attributes accurately and at a lower cost. In this follow-up, there is a need to know the size of the plots and the sampling intensity that provides estimated precision of production within a pre-established error limit compared to the forest census. What sampling intensity and plot form presents the best estimates of production, cover value and species richness when compared to the forest census, in the Eastern Amazon? As a hypothesis, we have: If the sampling intensity is equal to or greater than 5 % and using rectangular plots, the sampling will accurately estimate (with a sampling error of up to 20 %) the production, the coverage value, and the species richness of the ecosystem. With this, the objective was to analyze the sampling accuracy compared to the forest census in a native forest ecosystem in the Eastern Amazon, as a subsidy for potential areas of sustainable forest management. Eighteen sampling strategies were simulated with three intensities (2 %, 5 % and 10 %) and six sample unit dimensions (10m x 20m, 10m x 50m, 10m x 100m, 20m x 20m, 20 m x 50 m, 20 m x 100 m). The diameter of the inventory included trees with DBH ≥ 40 cm, from a list of 118 species considered commercial. The sampling system was simulated and based on the fixed area method and the two-stage systematic process. The simulation that showed the lowest trend in estimating the number of trees and basal area was strategy 14, with sampling intensity of 10% and sample units measuring 20m x 20m. Strategy 15 estimated the volume with the smallest error, with a sampling intensity of 10 % and sample units of 10 x 50 m. The coverage value identified the species Terminalia dichotoma, Osteophloeum platyspermum, Mezilaurus itauba, Dipteryx odorata and Tachigali paniculata as the ones that stood out the most and with the most stable relative errors in the 10 % intensity sampling. Strategy 18, with a sampling intensity of 10 % and a sample unit of 20 m x 100 m, presented the lowest error when estimating the number of species. Thus, the sampling strategies showed precision in the differentiated estimates according to the variable of interest to be evaluated, being an efficient technique to quantify and qualify the forest attributes, serving to identify areas subject to sustainable forest management plans in the Amazon.