Validating ranges, georeferenced records, and species distribution models, for vertebrate and plant species globally

Species range maps and georeferenced records of species occurrences supply complementary information about where species occur. I applied range maps and georeferenced records to examine intersection between range maps and the most certain and less certain locations of georeferenced records (i.e., coordinates can have an associated field of coordinate uncertainty in meters), intersection between an alternative range map product and georeferenced records, and areas of range maps relative to areas predicted by species distribution models from georeferenced records or range maps, with varying sample sizes. For all combined 3853 vertebrate and plant species, 25 % of species had at least 96.8 % intersection between georeferenced records of distance uncertainties <1 km and range maps, 50 % of species had at least 91.7 % intersection, and 75 % of species had at least 81.0 % intersection. The 25 % of species range maps with less than 80 % intersection may need reevaluation for mis-location, with omission and commission error. Similar results occurred for intersection with ranges of georeferenced records for distance uncertainties greater than 1 km or unknown, relative to uncertainties <1 km, and the Mammal Diversity Database (MDD) range maps, relative to International Union for Conservation of Nature (IUCN) range maps. Species distribution models based on small sample sizes overpredicted total area of occurrence, with about half of the predicted area outside of range maps. Range maps and georeferenced records triangulated, regardless of distance uncertainties. Range maps can help validate species distribution models or indicate issues of area overprediction and poor representation due to small samples, whereas georeferenced records detected potential anomalies in range maps. • Range maps and georeferenced records supply information about species. • 75 % of 3853 species had ≥81 % intersection between records and range maps. • Species distribution models based on small sample sizes overpredicted area. • Range maps and georeferenced records validated each other. • Range maps can help validate species distribution models.