I’m planning on exporting data from GBIF to test modelling species distributions; however, how accurate will they be if if the species is vulnerable or endangered, and the locations are obscured?
If this is correct, obscured locations are fake coordinates.
Depends on how fine-grained your modelling is. The obscured observations are represented as a random point in a 0.2x0.2 degree rectangle, which varies in size depending on how far from the equator it is. https://www.inaturalist.org/pages/geoprivacy
I believe that in some cases the GPS coords just don’t show up on GBIF for obscured data. They will give the uncertainty anyways and list it as around 28 km. I just tested it by pulling some GBIF data for bog turtles and looking at it, but I’d suggest just downloading a sample dataset from GBIF and seeing what it does with the iNat records you’re looking at.