When you make an iNat observation, you should imagine that iNat is asking you “what are the coordinates of your observation?”, not the more generic “where did you see it?”
While “Moquah Barrens State Natural Area” is a reasonable answer to the second question, it doesn’t work as an answer to the first question – “Moquah Barrens” is not a coordinate point, it is a polygon. Likewise the project “Moquah Barrens” is not a coordinate location, it is a project using the boundaries of Moquah Barrens State Natural Area.
However, iNat understands that people often don’t know the coordinates of their location, they only know a name or address, i.e. a text description of where they were. So to facilitate entering location information, iNat uses Google to try to get coordinates from a text description – this is called geocoding. For example, you can say you were at 1600 Pennsylvania Ave, and Google will tell iNat to use 38.8977149, 77.03660769999999 as the location:
If you say you were in Yellowstone National Park, this is what Google gives iNat:
Google’s geocoding isn’t perfect, and especially so for places outside populated areas. So anytime you use Google’s geocoding to map your location, you should double check that it looks appropriate and refine it yourself if possible. For example, after I got the Yellowstone result, I might refine the circle to just the area around Old Faithful by moving and resizing the circle:
When a project tries to determine if your observation is in it, it checks two things: are the coordinates within the project area, and is the accuracy circle within the bounding box of the project. The bounding box is the smallest rectangle that can contain the project, not the exact polygon of the project:
As stated above, your observations will be added to the Moquah Barrens project as long as the coordinates are within the bounds and the accuracy circle is within the bounding box (which for this project is essentially identical to the true bounds).