4.6 Conclusions
• The PAB framework enhanced the iSDM by helping in the selection of explanatory variables, as well as the ecological interpretation of the models. Nonetheless, in practice it needs high-quality data that are typically unavailable to fulfil this approach, especially regarding biotic interactions. In case of plant invasion, adequate data on the biotic component could be delivered by phytosociological relevés.
• The models reveals that two alien species with similar ecology and biology can vary considerably in their invasion pattern due to direct human interference. Therefore, the conservation options, derived from iSDM, should be focused on a particular species, not groups of species, even if they have similar ecology and are closely related taxonomically.
• The presence/absence data, in addition to their pre-eminence compared with opportunistic, presence-only data for species distribution modelling purposes, are still prone to some bias. Results of this study suggest that the bias is correlated with mistakenly reported species absence. The realistic confirmation of species absence can be assumed to need a high sampling effort for confirmed absence. Exclusion of the potentially undersampled plots increased the model performance; however, additional data are needed (e.g., richness of target species group).
• The employment of maps showing the historical distribution of invasive species enhanced the modelling by revealing potential differences in patterns of species spread into a region.