Assessing the coverage and quality of global tree occurrence data

Big data for tree species distributions: how big and how good?

Data resources used via GBIF : 36M species occurrences
Pinus taeda
Loblolly pine (Pinus taeda) by Andrew Dreelin via iNaturalist. Photo licensed under CC BY-NC 4.0.

Around 60,000 species of trees are known and thanks to researchers, museums, herbaria and citizen science programmes, a very large amount of primary biodiversity data exists on the world's tree populations.

This study sets out to assess the geographical coverage and quality of available tree occurrence data from five major aggregators, including GBIF. The authors develop a workflow to integrate and control data quality of species occurrences intended for species distribution modelling.

They identify 49,206 species in the data, representing almost 85 per cent of all known species. Among the 36 million occurrence records, 17 per cent are deemed high quality, however, only 15,140 species has enough high quality records to perform species distribution modelling. The majority of quality concerns among remaining records relate to duplicates and lacking coordinates.

The study finds that spatial coverage is high in Europe, North America and Australia, while large gaps still exist in key biodiverse regions such as South-East Asia and central Africa.

Link to original article

Serra-Diaz JM, Enquist BJ, Maitner B, Merow C and Svenning J-C (2017) Big data of tree species distributions: how big and how good? Forest Ecosystems. Springer Nature 4(1). Available at: