prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 3-4/2018 str. 29     <-- 29 -->        PDF

the total value of GDD does not change, remaining R2=74.1 %. The day of year is not a statistically significant variable in combination with the insolation. Therefore, for the early flushing form of Pedunculate Oak, we recommend the univariate model that features only the insolation as an independent variable: GDDearly = −27.651 + 0.539 * insolation.
For the intermediate flushing group, both variables (insolation and day of year) were statistically significant, and the total variability of GDD with this model is R2=90.6 %. Based on the univariate model for insolation, R2=89.1 %, which is a difference of only 1.5 %. Therefore, we conclude that the recommended model contains only the insolation variable: GDDinter = −48.084 + 0.690 * insolation.
For the late flushing group, both variables (insolation and day of year) were statistically significant, with a total explained GDD variability of R2=78.7 %. The R2 value for the univariate model for insolation is 70.8 %. Due to the R2 difference of 7.9 %, we recommend the use of the multivariate model containing the both variables (insolation and day of year): GDDlate = −237.839 + 0.559 * insolation + 2.479 * day of year.
A second-degree polynomial has proven to be the best model of the relationship between chilling and GDD (F (3,7) = 25.71, p<0.001), which indicates that if the chilling is approximately zero, the GDD group needs fewer forcing temperatures, and if the chilling units are above or below the value of ±40 units, then GDD group requires much more accumulated forcing (Figure 3). Although the parabola is shifted from the centre, the x parameter (0.1644) shows that this shift is not statistically significant (t = 0.54, p = 0.60). In contrast, the parameter of the quadratic term (0.012) is significant at the significance level of 9.6 % (t =1.92, p =0.096). Additionally, to break the dormancy, chilling is required to bein the range of ± 40 chilling units, and if the value is higher or lower than this number, the trees require significantly higher forcing temperatures.
Earlier investigations of Pedunculate Oak noticed the existence of different phenological forms that varied from early flushing (var. praecox) to late flushing (var. tardissima), leading to variations in the leaf unfolding between 1st March and 10th May (Mátyás 1967, Stojković 1991, Perić et al. 2000, Baliuckas et al. 2001, Wesolowski and Rowiński 2006). In this study, the phenoforms were properly separated into three groups – early, intermediate and late flushing types – based on GDD and three other variables (Table 1, Figure