I’ve been noticing this with such consistency that I can only assume this is a bug.
Over 2,000 observations with coordinates at the exact same intersection in San Francisco, however the location is clearly wrong.
I first started noticing observations of eastern species showing up in downtown San Francisco a little over a year ago. I had assumed this was users in downtown SF uploading observations with the wrong location, but if you actually look at the users submitting these observations, they are not located anywhere near SF. In most cases, this is the user’s only observation that shows up in downtown SF.
One theory I have is that observations being uploaded with no geographic coordinates are somehow defaulting to these coordinates in downtown SF.
That’s what I thought at first, too, but why would a user in eastern North America upload an observation with “San Francisco” as the location? Especially when that user has submitted other observations with their correct location.
I feel like I’ve seen staff post things with San Francisco as the location in examples; maybe these are tests? Or a glitch where somehow things default to that. Could be remembering wrong though.
Something similar seems to happen whenever location is just the name and not pinned to the map. I’ve seen the same for both “Mexico” and “Mexico City”…an inordinate number of observations from the exact same spot.
i’m not sure if it helps, but most (but not all) of the affected observations have either of the following coordinates (and most of the rest seem to be just north or south of these coordinates, along the same longitude):
I have also noticed that the location for observations will sometimes momentarily display San Francisco before loading everything fully. I assumed it had to do something with iNat starting in the Bay Area. So I think this is a bug.
So basically, when uploading from seek with no GPS coverage or poor accuracy it defaults to these coordinates in San Francisco? Why not just upload it with missing location? Weird.
who knows? but if it helps to put things in perspective, here are now 5.9 million observations created via Seek, and we’re looking at maybe 3000 observations that have these default coordinates. so the problem affects 50 per 100,000 of these observations. not nothing, but not super significant.
hey, this one is easy. when using seek to enter observation, the default location is san francisco. i know it, because it was extremely annoying having to manualy move myself from that distant land to central europe where i was standing. i have since deleted seek and use just inaturalist with gps coordinates, but i can imagine tons of people who did not bother or did not notice and just entered it with san fran location.
Ah yes - the intersection of Market and Van Ness, the source of much of SF county’s species count. I spend a lot of time searching observations in there and asking people to correct their locations.
It does seem to mostly be a Seek issue - I’m not sure how to filter for just website uploads, but the ios and android uploads seem pretty normal and seek has tons of obviously out of range species.
It would be better for such observations to be saved without a location, and ideally prompt the observer that the location is missing and how to fix that.
if folks are making the assumption that data in GBIF are 100% reliable and don’t require some basic scrutiny, then those folks face bigger problems than a few mislocated observations in the middle of San Francisco.
If they were obviously not taken at that location, you can mark them as “location not accurate” in the DQA and they will become casual. I suggest also leaving a comment for the observer to suggest that they correct their location.