SAMPLING USING THE BITTERLICH METHOD IN Pinus taeda L. PLANTATIONS WITH DISTINCT MANAGEMENT SYSTEMS

Abstract

The Bitterlich method was studied with multiple basal area factors (BAF) in 3 Pinus taeda L. plantations located in the Midwest of Santa Catarina, Brazil using different management regimes: without thinning (1344 trees.ha-1), with one thinning (789 trees.ha-1), and with two thinnings (475 trees.ha-1),. Using the parameters obtained from censuses, we sought to verify the data from different sample treatments for the variables V.ha-1, G.ha-1, N.ha-1 and “d” through a completely randomized design for each management regime and observing the precision obtained by the sampling error and the accuracy detected by the real error. It was observed that the variable “d” was less impacted by the change in the basal area factor (BAF), but the estimates of V.ha-1, G.ha-1 and N.ha-1 were more consistent (precision and accuracy) using factors equal to or less than 5, mainly in thinned areas. Even with no significant differences in any BAF in the 3 experiments, the sampling performed with BAF greater than 5 showed contrasting results in the managed areas depending on the BAF used. Sampling precision is influenced by BAFs, decreasing with the reduction in the number of trees counted within the Bitterlich sampling unit, a direct consequence of the BAF increase. Sampling errors were always overestimated, regardless of the dendrometric variable or area management. Even the largest differences between estimates and parameters being identified in BAF greater than or equal to 5, the experiment indicated the use of any BAF, with the largest requiring greater sampling intensity.