DIGITALNA ARHIVA ŠUMARSKOG LISTA
prilagođeno pretraživanje po punom tekstu



ŠUMARSKI LIST 7-8/1970 str. 20     <-- 20 -->        PDF

we applied logarithms of heights, whereby we achieved to some extent the homogeneity
of variances of logarithms of heights within individual diameter sub-classes.


For the height curves of definite strips (Nos. 1—5), the storeys (d = dominant,
n = auxiliary, p = understorey) and stand ages (Tabs. 2, and 2a) as well as for the
height curves of overall data (Tabs. 3 and 4) there were computed by means of the
method of least squares the regression coefficients (Ina, b), their error (slna, sb),
standard deviation around the smoothing line (slnh, d), ordinates of smoothing line
(Graphs 2—8), also 95 and 99"/o confidence limits. For the sake of better clarity of
graphs the confidence limits were not plotted on graphs.


On the ground of results of the regression analysis there was performed an
examination of regression functions and heigth curves.


By means of Bartellet´s test we examined first the homogeneity of variances,
variability of logarithms of heights in stands of different ages (Tabs. 4 and 5) and
proved the existence of significant difference between the standard deviations
around the smoothing line (sln]l, d) of overall heights and heights of individual
storeys. In like manner we proved that the heigth variabilities (s]llU, đ) differ
significantly also among the ages within the same forest community (Tabs. 2, 2a and
6). We are dealing with two forest communities, i. e. the forest of Pedunculate Oak
and Hornbeam (strip No. 1, Tabs. 2 and 2a), and the fooded forest of Pedunculate Oak
(strip No. 3, Tabs. 2 and 2a), also with four ages. By means of Fisher´s test it was
proved that there exist significant differences of variances (s2lllh, d) between the
forest communities within the same age for the overall heights (Tab. 7). However,
in the regression lines of dominant heights there exist no differences between the
variances (s*inh, a) of different forest communities within the same age (Tab. 8),
excepting the forest communities within the age of 90 years where this difference
is significant.


By means of an analysis of covariance (22) were examined the parameters of
regression functions. This test showed that there exist significant differences (Tab.
9) both as to the form (b) and position (In a) between the height curves of stands of
different ages within the same forest community. In other words, in the forest of
Pedunculate Oak and Hornbeam we cannot apply one height curve for all ages
(Graphs 5 and 6). The same rules also for the flooded forest of Pedunculate Oak
(Graphs 7 and 8). Regression coefficients increase with the stand age increasing also
in the height curves of overall heights (Tab. 4 and Graph. 2) and in the height
curves of individual storeys (Graphs 3 and 4) irrespective of the forest community,
also within a definite forest community. The height curves of older stands are
steeper.


Differences between the regression coefficients of the forest of Peducnulate Oak
and Hornbeam, and the flooded forest of Pedunculate Oak — within the same age —
are significant too (Tab. 10). In height curves of overall heights the mentioned
differences are significant both with respect to position and form, while in the height
curves of the dominant storey the differences are significant only with respect to
their position. For the 70 year-old stands the significance of differences is converse
namely, the dominant height curves differ in both parameters, while the height
curves of overall heights only in parameter In a. However, in general, we can say that
differences between the height curves of different forest communities but of the
same age are significant, and that these differences are more strongly pronounced as
to the position of height curve than on the slope. These differences, in fact, are the
differences of various forest site qualities, because the forest of Pedunculate Oak
and Hornbeam is found on good quality sites, while the flooded forest of Pedunculate
Oak is foiund on poor sites. It is interesting that heights of the flooded Pedunculate
Oak forest of 90 years old exceed significantly the hights of the forest of Pedunculate
Ooak and Hornbeam of the same age. Of course, this does not speak in favour of the
height as an indicator of site quality — in so far we agree that the stand age taken
from the working plan of the Lipovljani district is exact and that the flooded forest
of Pedunculate Oak occurs on sites of poorer quality. On the ground of the results
obtained through the analysis of covariance, we can conclude that the position of
the height curve is influenced by age and site quality, while its form is probably
influenced rather by the system of management than by site quality. The stands of
the same age but belonging to different forest communities are spatially not separated