Original paper

Sampling effort and factors influencing the precision of estimates of tree species abundance in a tropical forest stand

Cielo-Filho, Roque; Gneri, Mario; Antonio Martins, Fernando; Roberto

Abstract

Stand inventories are indispensable in community and population studies, diversity and conservation assessments, and pattern search, representing the first step towards understanding distribution and abundance variation of species in space. As species abundance descriptors are estimated through sampling, the precision of the estimates is important to assess data scope. In a 6.5-ha area of a semideciduous Atlantic forest, SE Brazil, we randomly located 100 plots of 10 × 10 m to sample trees with DBH ≥ 5 cm. We calculated the sampling error of estimates of density, frequency, dominance, and importance value index (IVI) for species with five or more adult individuals, and determined the number of plots necessary not to exceed sampling errors of 20%. Esenbeckia leiocarpa (Rutaceae), the most abundant species, was the only species for which sampling errors did not exceed 20%. The most appropriate criterion for evaluation of the sampling sufficiency for the inventory of the stand as a whole was the one based on the general sampling error of a set of the most abundant species. The estimates of density, frequency and IVI were influenced by the aggregation of individuals. The estimate of dominance had a greater influence of the basal area variation among individuals. Frequency had the greatest precision, dominance had the smallest, and density and IVI had intermediate precision.

Keywords

multiple plot sampling methodphytosociologysampling designsemideciduous atlantic forestspatial autocorrelationvegetation survey