Species-area relationships of plant communities and the possibility of predicting plant species diversity – a case study in South-Western Poland

The main aim of this study was to consider the possibility of predicting the number of plant species in areas occupied by many different habitat types. A very simple mathematical model was proposed for this purpose. It is based on two fundamental assumptions: first – every single type of community has its own species-area model relationship, second – the number of species common to various types of habitats allometrically depends on the number of habitats and on their quality. 
In order to test the proposed model, first the species-area relationships for single community types should be counted. Basic data were obtained from phytosociological tables published for SW Poland in 1960–2002. Each set of patches of a plant community, represented in one phytosociological table, was treated as one, compact habitat island of a size comparable to the joint acreage of the patches. The research covered all literature-documented plant communities from SW Poland – in all 750 phytosociological tables including 223 associations and plant communities. 
The data on 173 syntaxa compiled in 667 tables were used in the analysis – the remaining tables were represented by insufficient numbers of syntaxa (fewer than five), or by insufficient number of phytosociological releves (three or fewer). The species-area relationship models for 57 types of communities were counted this way.
The next step involved substituting the results of the single SPAR models in the previously proposed -diversity allometric model. The model was tested on 13 different-sized and 18 equal-sized areas in SW. Poland using GIS tools. In both cases the differences between the actual and predicted number of plant species does not exceed 12%. 
The consequences of the obtained results were discussed in the light of the main problems implied in the issue of species-area relationship.