An accidental finger brushing the mouse trackpad put my vote down as the first option, but I intended to vote for the last one, which is usually the case. The organisms that I photograph that are distant are either trees or birds. Birds are highly mobile, so the location I observe from is likely to be within the range of that bird’s likely travel. For trees, usually at least one of the photos in the set is taken right at the base, especially for individually notable trees, but if I’m just trying to tally species present in the immediate area from a trail, I’m not going to go through and try to manually guesstimate the location pin over to where the tree might have actually been located.
I do wish that in the upload process there was a way to specify which photo the observation’s location data should sync to (since the one taken at the base is usually not the one I want to be the primary view).
As a more pie-in-the-sky thing, I wish that gps-enabled cameras would use the focus distance and orientation data embedded in the exif to calculate the location of the object being photographed, not just the location of the camera itself. At longer distances at wider focal lengths (like for trees) it wouldn’t be terribly accurate, but probably still realistic to 30m accuracy. With long telephotos shooting birds it could probably be pretty accurate. It would really be an easy feature to implement, just a little extra code in the firmware, just not one for which there’s much demand. I can currently get the same result with any old camera if I take the bearing and a distance reading with a rangefinder from the spot where I take the picture (probably with much greater accuracy), but that means bringing two additional instruments into the equation when one would have sufficed.