The aim of this paper was to compare methods and sampling procedures applied to eucalyptus grown under different spatial arrangements in an integrated crop-livestock-forest system (ICLF). The study was carried out in Sinop county, Mato Grosso state, Brazil. Arrangements combining fixed and variable area methods to simple random and systematic sampling were tested. Precision, accuracy, efficiency, and diameter-distribution measurements were used for selection of the best sampling arrangement. The sampling intensity, sample unit optimal size and the relative efficiency have been changed depending on the spatial arrangement for 10% sampling error and 95% probability level. The fixed area method based on systematic sampling, as well as smaller sample units with eight to nine plants (48 m 2 to 54 m 2) were the most accurate; however, larger plots (192 m 2 to 216 m 2) were more efficient. All sample units size provided good estimates on the number of trees by diameter classes.

No abstract