Location Specificity and Underwater Observations

Hi All,

Many of my observations are taken in underwater environments. When I’m exploring close to home, I generally have an accurate idea of the specific geolocation of my observations, sometimes within several meters. I have many historical observations that I’ve avoided uploading because at the time taken I was traveling and/or less familiar with the area. Most of these were taken prior to my use of inat so I had no reason to log the specific location. Is there value in uploading historical images if I have no way of knowing a specific geotag - especially if many of these images are of common tropical reef organisms? In most cases I’m confident I could tag within a few miles of these offshore sites but that’s still potentially misleading and quite inaccurate for sessile organisms in particular. I welcome your opinion, thanks!

1 Like

It depends on HOW uncertain you are and maybe also on the commonality of the organism.
You can use the uncertainty circle around the location and make it as wide as you need to make sure that the real observation locality is for sure inculded. I did this all the time with my older pre-iNat-observations (done with those now). If you observed something interesting, it might also be interesting to have this observation tagged as “somehwere in the red sea” if you can´t be more precise. However, it might not be to interesting too know that you saw it somewhere in the Pacific ocean… unless it maybe should not occur in that ocean at all :-)

As I also use iNat as a collecting tool just for my own fun (Gotta catch them all ;-)) I just upload, be as precises as possible and enjoy and let others decide if it is useful for them as well or not.


I’m not sure about their usefulness scientifically, though I imagine that something uncommon could still be useful with an accuracy of several miles.

If you would enjoy it, go ahead and upload and just be conservative with the accuracy circle (ie erring on the side of too large to make sure it contains the true location).

One other reason these types of observations could be valuable is for training the AI. iNat generally has way fewer underwater observations, so increasing these is cool.


As long as its within the general area it probably doesn’t matter too much, still shows that there is a population of those sessile organisms in the area


Given that for global analysis the grid rez is often in the tens of kilometers scale my take is that this should be fine.

Ad a comment that the exact location is unsure, but that it is within X distance.

Also, look closely at maps of the region, especially bathymetry maps. Often you’ll see a feature and remember the location much more clearly. In cases like yours I often have a map open in a separate tab, or in a separate program entirely so that I can look around and get the location more accurate.


I think it should be fine. And you can create pinned locations in the web uploader to quickly put them in a general dive area.


Thanks all! I’ll start reviewing and hand picking some good photographs to upload. If nothing else, they’ll be food for the CV algorithm