The objective of this work was to determine the necessary sample size in order to estimate individual and populational production of Brazil nut trees (Bertholletia excelsa) in a native forest, and to evaluate different sampling systems. Data were obtained from a grove in the Itã region, Caracaraí county, RR, Brazil. Fruit and seed production from 239 trees was monitored from 2007 to 2010. Seven sampling procedures were tested based on the relationship of tree characteristics (diameter, morphological type, and canopy traits) with production, allocation type, and sample periodicity. To verify the precision and accuracy, 128 samples were taken from population, then it was calculated the standard error of mean, confidence interval, and the absolute error. To estimate tree production at 1% of probability, it is necessary to sample 97% of fruit, on average. For a 10% error limit, it is necessary to sample 33% of fruit. Continuous sampling, stratified by the diameter at breast height in six strata, and proportional allocation to the number of trees in the diameter class are the most suitable sampling procedures.
Bertholletia excelsa; native forest; forest inventory; nontimber forest products
