Besides the latency in the system that @cmcheatle mentioned above, there are a couple of other ways this could happen:
- It might have been a taxon split, meaning some observations moved to a new name, while others stayed under the old name. If you follow the link on the taxon change notice, you can see exactly what kind of change it was (and consider posting the link here, so we can all take a look).
- In the account settings, users have the ability to opt out of automatically accepting taxon changes. That has the potential to leave some observations identified under the old (but now inactivated) name.