Abstract
Understanding the relationship between biodiversity conservation and ecosystem services concepts is essential for evidence-based policy development. We used text mining augmented by topic modelling to analyse abstracts of 15 310 peer-reviewed papers (from 2000 to 2020). We identified nine major topics; "Research & Policy", "Urban and Spatial Planning", "Economics & Conservation", "Diversity & Plants", "Species & Climate change", "Agriculture", "Conservation and Distribution", "Carbon & Soil & Forestry", "Hydro-& Microbiology". The topic "Research & Policy" performed highly, considering number of publications and citation rate, while in the case of other topics, the "best" performances varied, depending on the indicator applied. Topics with human, policy or economic dimensions had higher performances than the ones with 'pure' biodiversity and science. Agriculture dominated over forestry and fishery sectors, while some elements of biodiversity and ecosystem services were under-represented. Text mining is a powerful tool to identify relations between research supply and policy demand.