Partner: Naturhistorisches Museum Wien, Natural History Museum Vienna (NHMW)
Use Case objective
Use Case 5 aims to assess and provide an alternative approach to the European habitat classification system by integrating vegetation species occurrences with Earth Observation data to predict their spatial distribution. It combined occurrence data from citizen science projects, herbarium collections, and scientific publications available through GBIF, along with topographic and climatic variables, and applied ensemble machine learning techniques to investigate the alignment between predicted species ranges based on real occurrences and official habitat maps. The expected outcome is an alternative, data-driven validation of phytosociological methods that enhances habitat mapping accuracy and the use of different data sources present in GBIF.
Application
UC5 proposes an alternative approach to the EUNIS habitat classification system by providing refined distribution models grounded in actual vegetation occurrences and key environmental drivers. These outputs can inform and support conservation planning, optimize sampling strategies, and guide museum digitalization efforts by highlighting ecologically relevant areas. The approach also illustrates the value of community science and herbarium data in enabling robust and reproducible habitat assessments. All generated maps and the accompanying modelling pipeline (R scripts) are designed to be FAIR-compliant and reusable across research, policy-making, and biodiversity monitoring applications.
Approach
A selection of habitats and their diagnostic species—starting with EUNIS habitat S22—were modelled across Europe using cleaned GBIF occurrence data and Earth Observation-derived environmental predictors. Species distributions were predicted using ensemble modelling techniques, integrating presence data with pseudo-absences generated via spatial buffering. Resulting species ranges were compared to EUNIS maps to identify discrepancies and improve habitat classification workflows.
