Dasgupta, SusmitaBlankespoor, BrianWheeler, David2024-06-272024-06-272024-06-27https://hdl.handle.net/10986/41788This paper builds on recent advances in machine-based pattern recognition to estimate species occurrence maps, using georeferenced open-source data from the Global Biodiversity Information Facility. With currently available data, the estimation algorithm has produced maps for more than 600,000 vertebrates, invertebrates, other animals, vascular plants, and fungi. The algorithm is designed for rapid map updates and estimation of new maps with continued increases in Global Biodiversity Information Facility occurrence reports. The paper compares the algorithm-produced maps with species-matched sets of expert maps for mammals, ants, and vascular plants. Using comparative species density counts in a high-resolution grid, it finds close similarity in global distribution patterns. It also traces regional differences to technical differences in estimation methods or cases where the boundaries of existing expert maps could be revised to reflect species-level patterns in Global Biodiversity Information Facility reports. The paper uses the estimated Global Biodiversity Information Facility maps to explore the global distributions of endemic species and species whose small occurrence regions increase their extinction risks. It finds a high overall incidence of endemism, with significant variations across major species groups. It also identifies about 118,000 species whose small occurrence regions create significant extinction risks. For both endemic and small-occurrence region species, the paper finds patterns of spatial clustering that identify candidate areas for cost-effective protection.en-USCC BY 3.0 IGOCONSERVATION PLANNINGGLOBAL BIODIVERSITYSPECIES' OCCURRENCE REGIONENDEMIC AND SMALL-OCCURRENCE REGIONKUNNING-MONTREAL GLOBAL BIODIVERSITY FRAMEWORKLIFE BELOW WATERSDG 14LIFE ON LANDSDG 15Working PaperWorld Bank10.1596/1813-9450-10821