EFFECTIVENESS OF SAMPLING METHODS FOR THINNED AND NON-THINNED Pinus taeda L. PLANTATIONS IN THE MOUNTAIN REGION OF SANTA CATARINA, BRAZIL

In Brazil, most forest inventories by sampling use the fixed-area method, even though there are other alternatives methods developed since the 1940s. Research demonstrating the effectiveness of variable-area sampling methods is scarce, especially in forest plantations that advocate thinning throughout the production cycle. This study aimed to test different sampling methods of fixed (400 m² circular) and variable (Bitterlich and Strand) areas in a forest inventory in 9-year-old Pinus taeda L. plantations with 9 years without thinning and 14 years with thinning, comparing with the total enumeration of areas. The results obtained from the variables Number of Trees per hectare (N ha-1), Basal Area per hectare (G ha-1) and Volume per hectare (V ha-1) were submitted to a test of average comparison following a completely randomized design for each plantation and the differences observed were detected with the Scott-Knott test at 95% probability. It was evidenced that the Strand method presented greater accuracy for estimating all analyzed variables, especially in the area with thinning, however, it is necessary to increase the sampling intensity to guarantee precision in all cases. The fixed area method showed results consistent with the census for N ha-1 and presented overestimates in the area without thinning and underestimations in the area with thinning for Gha-1 and Vha-1.