Central-European vegetation types and their optima along successional gradient

Although the identification of plant communities is the basic language of communication, studies that focus on the classification of vegetation in successional series are rather rare, mainly because it is difficult to identify different types of vegetation. Thanks to formalized algorithms of machine learning, we were able to assign some of vegetation plots stored in a Database of Successional Series (DaSS) to alliances in the vegetation classification system. Of the samples in DaSS 67.4% were classified into 96 vegetation alliances. Classification of the seral stages was then used to predict optima and intervals of occurrence of 33 main types of vegetation in the first 70 years from the onset of succession. In accordance with general expectations, main types of vegetation were arranged at the time-scale from ruderal and segetal vegetation, across grasslands to shrubby and forest vegetation. Successional optima of particular units of vegetation can be used to roughly predict the successional changes at human-disturbed sites in central Europe.