Testing ecological hypotheses using ‘synthetic’ datasets

Large data networks provide researchers with new means of addressing a large number of macroecological questions without having to collect new data. This study reviews common methods and principles, identifies bottlenecks and provides clear recommendations for using these approaches. The authors created a case study on a pine-marsh food-web structure based on interactions from a variety of database sources and species distribution models created using GBIF-mediated occurrences. They demonstrate that synthetic datasets like these can support large-scale qualitative predictions while identifying gaps in our knowledge of biological systems. The results describe methods that can aid in identifying high-priority areas for fieldwork.