The aim of this work was to compare the estimated values obtained by sampling with the actual values recorded in the Forest Census (inventory 100%) of a forest area in the Amazon. A systematic sampling was simulated, attending the requirements of IBAMA, with 2-ha sample units and sampling intensity of 14% in a 1,000-ha area, which was subjected to a forestry census. The census and simulation of forest inventory considered only the population with DBH > 40 cm and a list of 60 species which were identified as the population of trees considered of commercial value. It was evaluated the deviations of abundance, basal area and volume per hectare of sampling in relation to population. Deviations were evaluated with respect to individual species and species grouped according to the classification of price used by the Brazilian Forest Service. Deviations were also evaluated by diameter class, by group, and the total population. Even with sampling intensity of 14%, six species (10%) were not recorded in the sample, resulting in deviations from more than 614% to more and 100% to less in relation to abundance, basal area and volume.
Forest inventory; Sampling intensity; Accuracy
