This study was accomplished with the following objectives: 1) to evaluate the acuracy
of non- segmented polynomial models to estimate diameters and volumes along the
stem of Mimosa scabrella, by diameter class, by age, and by the total of the data; 2) to
if verify the if best model to estimate diameters is also the best to estimate volumes; 3)
to proceed the validation of the models of better acuracy for volume estimate with
trees cubed through the technique of water displacement (xilometer); 4) evaluation of
the taper equantions to estimate volumes of multiple products of Mimosa scabrella
wood; 5) to test the possibility of grouping the models in the different age classes
through the identity test. The data base was composed by 174 trees of Mymosa
scabrella, where 121 trees were cubed by the method of Hohenadl with ten sections
and 53 trees by the method of water displacement (xilometer). All sample trees came
from the Meteropolitan Area of Curitiba/PR. The 121 trees were divided in 5 diameter
classes and 4 age classes. It was proceeded the adjustment of three non-segmented
taper models for each diameter class, for age class and for the total group of data. The
tested non-segmented models were those from Schöepfer (1966), Hradetzky (1976)
and Kozak, Munro and Smith (1969). The acuracy of the tested models for the
estimates was evaluated through the following statistics: graphic analysis of residual
adjusted R2, standard error of estimate in percentage (Syx%) and graphic analysis of
average profiles, just used to verify if the fittings are satisfactory in a general way.
Complementary the following statistics were calculated: deviation (D), standard
deviation of the differences (SD), summ of squares of relative residues (SQRR) and
percentage of residues (RP) in order to have a detailed analysis of the performance of
estimates along of the whole stem. Starting from these complement sy statistics a
ranking was performed to detect the model that propitiated more accurated estimates
of diameter sand volumes along the stem. The volumes of the 53 xilometered trees
were then estimated through the equation of better performance for the grouped data
by diameter class and by age, as for the total of trees. The estimated volumes of these
53 trees were compared with their respective real volumes through the c2 estatistic.
The assortment was done for individual ages and for grouped ages through the better
fittings equations for volumes estimates, and later on extrapolated to hectare. In the
attempt of grouping of the classes of age the identity test was used, facilitating to do
11 combinations. After the analysis it was concluded that the model from Hradetzky
(1976) was the best for estimating diameters will as estimating volumes, in almost all
diameter classes and ages. It was also the best for the total data set. The model from
Kozak et al. (1969) presented bias estimating diameters and volumes for all classes
and for the total data set. The validation test proved that taper equations of best
performance for diameter estimation along the estem ware also the best for volume
estimation. Therefore they can be used to do the assortment. In general the obtained
assortments presented great variation among and inside of the ages, caused by the
oscillation of the reason height/diameter The identity test for the three tested models
showed that they can be grouped for the age classes 3 and 4. the age classes 1 and 2
only can be grouped for the 5º degree and for the Kozak models.
