The BERN model (Bioindication for Ecosystem Regeneration towards Natural conditions) was developed to integrate ecological cause-effect relationships into studies on environmental status assessment and forecasting. Common approaches such as ordination (Roberts 1986) or bioindication based on individual species like PROPS or MultiMOVE model are of limited use because it is not possible to predict the occurrence of a plant species on the basis of site factors, since the competitive influences cannot be determined in advance according to current knowledge. Therefore, the BERN model takes into account the structure of plant communities with the abundance and cover rates of species in the competitive equilibrium of pristine plant communities as a reference for determining anthropogenically induced changes. Qualitative knowledge on the relationship between site types and vegetation communities is widely available, as can be seen from the extensive phytosociological publications. The aim of the BERN model is to make part of this knowledge available in numerical form for computer-based ecosystem modelling. The task of the BERN database is to document reference data for the occurrence of plant species in competitive equilibrium with each other and in equilibrium with the site factors. The BERN model combines the fundamental species niches that mainly form the competitively stable structure of a pristine plant community in order to determine the realised multifactorial niche of this community. The community database includes 887 central European plant communities and links to their diagnostically defining species composition. The database defines the niche of 2210 central European plant species for the soil properties pH, base saturation, carbon to nitrogen ratio and wetness index and the climatic properties continentality, length of vegetation period and climatic water balance. Another difference to all the other models mentioned above is the BERN database with historical recordings at more or less undisturbed sites. The BERN database now contains so many vegetation relevés that the database can be considered representative at least for Central Europe. The methodology of creating the BERN database and the BERN model is documented and applications are demonstrated with examples.