engleski

Spatial prediction of ground beetles in beech forests on karst

Beech (Fagus sylvatica L.) is most widespread tree species in Croatia, which constitutes several forest communities on karst. In Croatia, spatial arrangements of these forests is different in Dinaric part where they creates continuous matrix, while in Pannonian part it is fragmented to isolated mountains. Plant diversity of these forests is influenced mainly by climatic conditions. Here, we have used field data on ground beetles from several mountains across Croatia, and data on five environmental variables for the whole territory of Croatia (elevation, annual precipitation, mean-, minimum- and maximum annual temperature). Using Maxent software we have developed predictive distribution models for 24 species which had at least 10 recorded localities. Exemption to this was Calosoma sycophanta with eight records. Although, as expected, all five environmental variables were correlated, developed models has shown patterns according to spatial distribution of species that can be divided into three groups. Two biggest groups were for species distributed all over beech areal, and that with species more restricted to Dinaric part of beech areal. Distributions of species from former group were mostly influenced by elevation and maximum temperature, while those from latter group were influenced by precipitation and elevation or maximum temperature. Third group consisted of just two species ; Abax parallelus and Carabus ullrichi was limited to the Pannonian part, and predicted mostly with minimum temperature, followed by maximum temperature and precipitation. Five species that did not show clear membership with neither of three groups were all distributed across complete areal with exemption of Leistus spinibarbis which was restricted to Dinaric part. The highest percentage of single environmental variable’s contribution to the predictive models were those of precipitation for Molops ovipennis, and elevation for Leistus spinibarbis, 90% and 89.4% respectively. Majority of species had over 85% of contribution to the models obtained with just two environmental variables. Models without dominant environmental variable, that has four out of five predictive variables accounted for at least with 10% of contribution was developed for Abax carinatus, A. parallelus and Carabus ullrichi.

Carabidae; Fagus; Croatia; MAXENT

nije evidentirano