We have this page on the obo rendering, where we have tried to make this as transparent as possible, but always room for improvement:
http://obofoundry.org/ontology/ncbitaxon
I think we at least need to make clear on this page that there is a separate bioportal/UMLS rendering. And we should document differences in content, PURLs, and release cycle.
We also need to inject metadata into the header that links to this page. Someone landing here:
https://www.ebi.ac.uk/ols/ontologies/ncbitaxon
is none the wiser that they are looking at a specific OBO rendition
I like how on the bioportal page it's (more or less) clear that this comes from UMLS
http://bioportal.bioontology.org/ontologies/NCBITAXON
However, there is a lot that is undocumented on both sides. For example, Bioportal/UMLS excludes non-organismal metagenome entries, excludes the fake 'root' class; OBO has a faithful rendering of every NCBI taxonomy ID. There are pros and cons of both approaches.