Are you aware that iNat has two different lists of locations? My understanding is that they’re defined somewhat along these lines.
-
If you use the Location field in the Identify dialog, this uses a list of places managed by iNat: https://www.inaturalist.org/places
-
If you use the Location field at the top of the Explore page, this supplements the iNat places with a list of locations sourced from Google. It’s pretty common to encounter places that are defined differently than one might expect. Any fixes to these place definitions need to be taken up through Google’s processes. There are several related discussions on the forum, such as this one: https://forum.inaturalist.org/t/how-is-accuracy-of-google-maps-locations-determined-and-can-it-be-altered/40018/6
[I may have this slightly wrong, so maybe someone can provide a clearer definition or link.]
My understanding is that iNat’s places (item 1) generally encompass countries, plus the next two lower subdivisions (e.g. states and counties), plus custom-defined places set up by iNat users. So iNat places typically don’t include individual cities. In the case of Deer Park, TX, iNat has a defined place for Harris County.
Also, those “custom-defined places set up by iNat users” include a whole lot of poorly-defined places set up by early iNat users when the platform provided a visual place editor (think of a basic rectangle that vaguely encompasses the user’s home town).
It appears that if you enter a place name that isn’t an exact match, iNat will choose whatever appears to be the best text match, which often will be one of those poorly defined custom places. I think this is behind the mismatches to Deer Park Prairie, Vancouver Coastal Waters, San Jose State University, Los Gatos Creek County Park, etc.