Towards a global insect trait database

Useful knowledge for the conservation of biodiversity is still lacking and the current population trends for most insect species are unknown. Identifying ecological or functional trait characteristics of species that are the ultimate causes of extinction may be extremely useful to circumvent the lack of knowledge on species trends. A recent major work by our team found that habitat range and speed of life traits are in fact potential universal predictors of extinction risk for all terrestrial taxa, from vertebrates to invertebrates and plants. And yet, insect trait databases that could help predict extinctions are still fragmented taxonomically and geographically. The area of machine learning has seen major breakthroughs in the last years, with computers being able to execute functions recently thought possible only to human reasoning in extremely efficient ways. I will present current efforts to use machine learning to mobilize and make accessible under the FAIR principles multiple insect traits and a database that will at first try to cover traits known to be related with extinction risk and ecosystem service provision. Together with increasingly efficient ways for trait imputation we open new possibilities for conservation, functional diversity and ecosystem service research using multiple insect taxa.