prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 5-6/2022 str. 53     <-- 53 -->        PDF

revealed the optimal model for each species (Table 2, Fig. 2). Only the P/A ratio was included in both final models; however, the effect of this shape index was opposite and statistically significant only for P. serotina.
Our results suggest that the probability of P. serotina occurring in forest patches decreased as the P/A ratio, estimated nutrients and light indicator values increased. The effects of the P/A ratio and nutrients were statistically significant. Surprisingly, the probability of P. serotina being present also decreased if walking paths existed in a forest; however, this effect was only marginally significant (Table 2).
The probability of I. parviflora occurring in a forest patch increased with the increasing presence of roads in the buffer zone surrounding the patch, increasing P/A ratio, higher estimated soil moisture indicator value and with the presence of waste in the forest patch. However, only waste presence had a significant effect. The probability of I. parviflora being present in a patch was 6.54 times greater if there was waste detected, compared to patches without waste (Table 2).
The lowland forests in this study patches exhibited the general pattern observed for European woodlands (Wagner et al., 2017): I. parviflora, which is the most common alien