Would it be possible to add some angles to the corresponding polygon so that it follows more closely the course of the Iguazú River which forms the border?
Same issue for me in two observations made on the Brazilian side of the Parque Estadual do Turvo in Derrubadas - RS - Brazil. Their location appear to be right, but if I search my observations for anything in Argentina i get those 2 as the result.
The egret was on our side of the river. I can understand this record being shown on both sides if the accuracy of the geotag overlaps to the other side, but I don’t get it why the beetle appears there as it was nearly 500m away from the border.
I guess I don’t really understand how country is “assigned” to the observation, as my observations were precisely geotagged, and the exact locations are clearly on the Namibian side of the border and not the Angolan side!
iNaturalist simplifies complex borders in order to keep the site usable. If we used complex boundaries for borders at follow rivers, as in this case, it would make the site too slow to use. So here is the boundary for Nambia in the Rundu area:
You can see it doesn’t follow the river too precisely, thus some observations made on the Namibia side will end up in Angola, at least according to iNat.
What you can do is put the point of your observation within the Namibia boundary and increase the accuracy circle around it to encompass the actual location. Not ideal, but it should work.
Thank you for this very clear explanation! Much appreciated!
Honestly, it doesn’t really bother me as to which country these observations end up in. From a scientific perspective, I’d rather they have the correct GPS coordinates associated with them (especially since I went to the effort of recording them!). All things considered, I’ll probably just take the path of least resistance and leave them as is
I know there are some algorithms out there for taking large, complex borders and pruning them down to much smaller sets of points while deviating as little as possible from the original. It might be worth experimenting with running some GADM shape-files through one or more of these algorithms a few times with different parameters, to see if you can get a result that leaves less area on the wrong side of the border.