This is somewhat a “bug report” and somewhat a “feature request,” so I think it counts as a general post. It has come to my attention that RG observations at the species complex/group level show as species records (nominally) on GBIF, which is a big issue. For example, take this observation of “Pheidole az05,” an undescribed species that fits firmly into the currently defined Pheidole pilifera group as a relative of Pheidole tepicana; on GBIF, it’s listed as Pheidole pilifera, which is very incorrect. This not only happens with observations at the group level because they represent undescribed species, but also on those that likely represent described species but aren’t identifiable to that level. I don’t think all of this is necessarily a “bug,” but rather a miscommunication between iNaturalist and GBIF’s databases. I’m wondering what can be done, and whether or not this is an issue only on iNaturalist’s side, or if GBIF would have to make accommodations to accurately represent these records. If groups are not a possibility on GBIF, I think that iNaturalist needs to be sending these records to GBIF at the genus level, because the current way is giving inaccurate data.
GBIF has an open issue for this already. I suppose you could ask iNat to send rank complex observations at genus instead. Unclear which organization is likely to change quicker 
URLs (aka web addresses) of any relevant observations or pages:
https://www.inaturalist.org/observations/57986223
https://www.gbif.org/occurrence/2883169445
The above mentioned RG observation of a Tachysphex terminatus (species complex) was exported to GBIF as T. terminatus (species). Species complexes and species are not the same. iNaturalist should make sure that identifications are exported to GBIF at the correct taxonomic level, e.g. species to species and species complexes to species complexes. I am not an expert on the structure of GBIF, which may not have a species complex category in their framework. In this case, the ID would have to be exported at the next higher taxonomic rank that is available (genus in this case).
I also noticed that the species complex identification is only supported by one identifier. Strictly speaking, only the genus is research grade in this case (supported by three identifiers), but in GBIF it shows up as a species record. I became aware of the problem because T. terminatus is not known to occur in BC, but now it has a record on GBIF that is actually completely unsupported.
I unmarked it from RG, because clearly user marked it before the complex id, and the second complex id is needed, there’s a clear problem because community taxon was exactly a genus, so why did species got to GBIF? I marked many things in a cimilar situation as this one, does it mean they also got to GBIF as species with 1 sp 1 genus ids?
Thanks, that’s exactly the same problem. I did not know that there was a thread already.
Before there is a permanent solution in place, iNaturalist should stop exporting species complex IDs to GBIF (or export them at genus level).
(Moved the four posts above to this topic to consolidate the discussion)
After some chasing up by @ydobon, this issue has been resolved. iNat records ID’ed to a complex now map to the genus in GBIF
The issue may be resolved–record to highest certain taxonomic level, i.e., genus, for accuracy when species complex is misinterpreted by GBIF as species–yet information is ‘lost’, namely, the finer level distinction that occurs in the original iNat record. GBIF does have the means to store that finer taxonomic information, but is not managed in same way (not as a name) as we can on iNat, as the GBIF taxonomic backbone does not allow for species complexes. The DarwinCore standard has several other data fields apart from the principal one of scientificName that could report/record this finer level distinction, e.g., taxonRemarks, verbatimIdentification ? The feature request then could be ‘If not an exact match in GBIF backbone, record to next-highest available taxon level in GBIF, and report original identification in dwc:verbatimIdentification’. Thus, the iNat species complex information would be conserved in GBIF records. https://dwc.tdwg.org/terms/#dwc:verbatimIdentification
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.