Amostragem para inventário florestal em sistemas silvipastoris

摘要

Forest inventory, according to many sampling processes, produces production estimates for the entire forest stand. In this work, it was tested many sampling processes in a silvopastoral system with band arrangements of Eucalyptus grandis in Umuarama, northeastern Paraná. The distribution of E. grandis bands were plotted in a map in aerial scale, in which it was applied some different sampling techniques: simple random sampling (with parallel rectangular plots ALS-RPA, perpendicular to the bands ALS-RPE, and circular plots ALS-C), and systematic sampling (with 5%, 10% and 20% of intensity, respectively, AS-5, AS-10 and AS-20). For comparison among sampling processes, the sampling error (E%) was considered, according to which it was verified to be the best methodology. The observed results showed that the greatest E% were obtained in the simple random sampling: ALS-C (61.64%), ALS-RPA (47.30%) and ALS-RPE (31.19%). For the systematic sampling, it was observed smaller E% values: AS-10 (6.05%), AS-20 (7.04%) and AS-5 (12.49%). The systematic sampling is efficient in forest inventory in silvopastoral system due to its low E%.