Some of my observations show up in locations that aren’t the location I’ve assigned them, and it is upsetting to be, as I’ve taken the time to attribute a precise location, and this will mislead any other user viewing the observations.
The example provided (URL below) is of a leafminer viewed on a live cottonwood; using the satellite imagery, I can see the cottonwood, and thus positioned the location marker on the cottonwood, yet afterwards, while viewing the observation (using the explore feature (website and app), and in the observation page mini map) I noticed that the observation marker is instead located within the soccer field, near the soccer goal.
I’ve noticed the same on many observations that can be precisely traced with landmarks which can be viewed by satellite imagery (e.g. corner of buildings, fences, canopy of isolated tree, etc.). This wouldn’t be such an issue if I hadn’t taken the time to manually position the location markers for these; also, some of my observations featuring a tree end up having their markers beside said tree, or street side observations have their markers in the middle of the street. My girlfriend also had this bug happen, where many of her terrestrial observations’ manually positioned markers were displaced into a lake.
you should never assume that satellite images or even street maps are particularly precise. there will always be some positioning error, and you can see how it seems like the satellite images shift a bit at different snapshots in time:
I had an interesting thing today manually placing something that had failed to geolocate on IOS - it did have an accuracy, and the accuracy changed depending on how far I zoomed in before placing the observation. I decided +/- 24m was about right for how accurate I probably was, so left it at that zoom, but I could have gone narrower or broader. It’s curious.
But yeah GIS nonsense is definitely nonsense, especially since presumably we’re all on WGS84 and while that’s the best datum on a global average it’s unlikely to be the best datum for your location, so at a minimum you’re probably off by +/-5m or so regardless. (The earth is not a perfect ellipsoid, after all, and all datums are based on one perfect ellipsoid or other…)
this is how the mobile apps record the accuracy value for manually positioned observations. it’s not how i would have designed it, but it’s not a new behavior.
I just hadn’t noticed that was how it worked before; it felt relevant to the lack of an accuracy measurement, I suppose, depending on whether one wishes to pick the number. Mostly I can confirm that perfect accuracy is +/- impossible with a GIS system regardless.
lack of accuracy value on an observation usually means the location came from the file metadata of an image taken in an Android phone’s camera app or geotagged using a track.
but, generally, the values (or lack of value) are meaningless in the grand scheme of things because there’s no consistency in how they are recorded across observations.
