PREDICTS (Projecting Responses of Ecological Diversity In Changing Terrestrial Systems) is a statistical model that represents many more species than previous models. It makes the most comprehensive predictions to date about past and future structure and diversity of ecological communities. The project:

• gathers existing data that describe the diversity and composition of biological communities at local scale around the world. By October 2013 the project team had compiled just over a million biodiversity measurements, representing more than 24,000 species in 67 countries.
• incorporates these data into a statistical model to predict the responses of ecological communities to habitat loss and degradation
• provides information to ongoing policy processes such as the Global Biodiversity Outlook.

The PREDICTS database is a globally and taxonomically comprehensive database of site-level measures of biodiversity at thousands of sites around the world. The 2016 release of the database contained more than 3.2 million records from more than 26,000 sites in 94 countries, representing 47,044 species. The taxonomic distribution of taxa in the database is in rough proportion to the numbers of described species in major taxonomic groups (Hudson et al., 2016). 

This database is used to statistically model how total abundance of organisms and compositional similarity respond to land use and related pressures. The PREDICTS team (including NHM, UNEP-WCMC, UCL) combines these models with spatio-temporal projections of explanatory variables (at 1km spatial resolution). Model results have been combined with HYDE data to hindcast BII values for the years 900 to 2015, and with five Shared Socioeconomic Pathways (SSPs) to project biodiversity futures for the years 2015-2100. The 1km resolution dataset can be used to calculate mean BII (weighted by factors such as net primary productivity or species richness) can be at the country, subregion, inter-region and global level.