DIGITALNA ARHIVA ŠUMARSKOG LISTA
prilagođeno pretraživanje po punom tekstu



ŠUMARSKI LIST 3-4/2018 str. 52     <-- 52 -->        PDF

Capture probability (p) was dependent on the periods, however the estimated monthly values of this parameter did not differ significantly when comparing the investigated forest habitats except in August. In this month the largest capture probability was estimated by the accepted model in case of the reforested habitat which was significantly higher than the capture probability value of the other two habitats due to the non-overlapping 95% confidence intervals. Similarly, the estimation showed significant difference between the protected forest and the habitat under forest management (Fig 4.A).
However, despite the difference in August, the distribution of estimated capture probability was not significant between forest habitats (one-way ANOVA: F = 0.498; n.s.). Regarding the seasonal change of capture probability, there was no significant difference when comparing the primary (monthly) capture periods in the protected mature and old-growth forest habitat while the estimated p value was significantly higher in September than in the previous two months in this sampling plot. In case of the Repaš forest in Croatia, the estimated capture probability reduced from July to September, although the decrease of this parameter was considered to be significant between July and September but not for consecutive periods (see the overlap of 95% confidence intervals of estimation). Capture probability increased by October in this area, too. Despite the high confidence interval of the estimated value in October, a significant increase of capture probability was characterized by the accepted POPAN model between September and October (Fig. 4.A).
The POPAN models calculated the size of population in each sampling period in each habitat as derived parameters. Based on the best candidate model the population size of bank vole in the protected forest habitat and the habitat under forest management changed periodically along sampling periods. However, in case of these habitats, the difference of population size was not significant between months due to the overlap of 95% confidence interval. The temporal change of abundance was similar in the reforested habitat but population size in July was significantly higher than in the further months and the difference of abundance in September and October was also significant (see the non-overlapping 95% confidence intervals) (Fig. 4.B). When comparing the calculated population size of the three habitats in a given sampling period, abundance did not differ significantly in July while in the another three months population size of bank voles in the protected forest habitat was significantly higher than in the reforested habitat. Based on the overlap of confidence intervals, abundance did not differ significantly between the reforested habitat and the managed forest (Fig. 4.B).
According to the second best candidate model where the estimate of population size (N) depended among the different habitats, the estimated number of super-population in the protected forest was significantly higher than in the reforested habitat. Based on the overlap of confidence intervals, super-population size did not differ significantly between the protected and the managed forest as well as between the reforested area and the habitat under forestry management. This result showed that the protected forest stand was the most suitable habitat for the bank vole (Fig. 5).