prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 1-2/2017 str. 12     <-- 12 -->        PDF

Linear regression analyses were used to test all the variables in relation to budburst dates (Table 1). The most significant values of R2 (INS_NOV, INS_DEC, INS_JAN, PREC_DEC) are presented graphically in Figure 2.
Multivariate regression analysis was used to find the best set of variables which would describe budburst dates. Using the leap function, the three best subsets for each subset parameter were tested, and then ranked them according to the R2 criterion (Figure 3). As seen from the graph, five variables had to be included for the lowest R2 value of 0.99 (TEMP_NOV, TEMP_JAN, INS_DEC, INS_JAN and PREC_DEC), whereas eight variables had to be included for the highest value (all except INS_DEC).
The last step in testing the impact of environmental variables on the beginning of budburst also involved multivariate regression analysis with stepwise method and step function. The criterion used was the Akaike information criterion (AIC, lower values – more significant impacts). The results of stepwise multivariate regression analysis are shown in Table 2.
The best model for describing the beginning of budburst according to the AIC criterion is the fifth model.
Temperature – Temperatura
Temperature is the most widely investigated climatological parameter (environmental driver) in terms of phenological modelling. Pinto et al. (2011) used the example of Quercus ilex and Quercus suber to study the relationship between mean daily temperatures and budburst dates. Along with all the other factors, it was temperature that proved to be the most important. In contrast to the mentioned study, this work uses the example of narrow-leaved ash to sum mean daily temperature values during the four years of research and from three different starting dates. Temperature values did not have an important impact on budburst dates. This fact serves to prove that there is no uniform prediction model and there are no identical requirements that affect the beginning of leaf unfolding in forest trees.