The use of sampling techniques in forest inventories generates information regarding
natural resources, contributing to the decision making of managers in the public and
private spheres. Considering the importance of information generated by the use of
sampling techniques, the objective of this study was to assess the influence of
different plot sizes and sampling intensities to estimate the stock and
phytosociological parameters. The studied area corresponds to a Mixed
Ombrophylous Forest remnant, of 15.2 ha, located in the Campus III – Jardim
Botânico, which belongs to the Federal University of Paraná. The methodological
approach of this study consisted initially of the georeferencing of the study area and
a census, of all trees with breast height diameter (DBH) greater than 10 cm. The
collected data were included in a geodatabase. After this step 20 sampling variations
of fixed area were defined to perform the inventory simulations. This process
considered the combination of different plot sizes (200sq.m, 400sq.m, 500sq.m and
1000sq.m) with different sampling intensities (2%, 5%, 10% and 15%). The results
indicated that the area contains 121 tree species, distributed in 47 families and 80
genera. In terms of species diversity the predominant family is Myrtaceae. The
average forest basal area is 22.05sq.m/ha and the average number of trees is
601N/ha. The comparison of the census results with those obtained through the
sampling simulations indicated that the sampling technique is quite reliable when the
forest is considered as a whole. However, when the results are evaluated at the
species level, they show a strong deviation in relation to the parametric value. In
general an increase in sampling intensity provided an improvement of the estimations
at species level. However it was observed that for plots of 200sq.m the variable of
interest presented a highest value. The opposite was observed in larger plots. Plots
of 1000sq.m tend to capture more accurately the variation of the spatial distribution of
the variable of interest. Thus the obtained results are more reliable than those from
200sq.m plots with the same sampling intensity. In addition, plots of 1000sq.m were
more precise for the estimation of the species cover index. The results corroborate
with the hypothesis that to estimate the stock of a forest and phytosociological
indexes larger plots present a better performance, particularly for species of great
representativity in the forest. The only situation in which smaller plots might perform
better is to estimate indexes related to rare species (less than one individual per
hectare). Based on this it is recommended that for estimations regarding stock and
phytosociological indexes, plots of at least 1000sq.m should be used and a sampling
intensity as high as time and financial resources allow.
Keywords: Forest inventory, sampling method, precision, phytosociology, spatial
distribution, diametric distribution.
