I've started verifying records and found a few iNaturalist records where the recision recrded for the record is so big the record grid reference is too imprecise to be worth having eg 'SD'. Should I accept the record, or reject it because the grid reference is inadequate, even though the iNaturalist record is good?
I've also noticed that many records from BirdTrack don't show an accurate grid reference but they seem to show a dot in a sensible location.
For some of those records that (ones where the observer has obscured the location on iNat) that's the actual precision that has come across. Only option there I find is to either reject or to go to the record on iNat and ask if the observer will remove the obscuring - it should then eventually come back to iRecord with a more precise location.
For some the data has actually come across with a pretty accurage location, but it is in Lat,Long format with an uncertainty radius (so essentially a circle). To generate a grid reference iNat then takes the smallest sensible OS Grid reference that entirely covers that circle. Often that works, but theoretically a record could be submitted with a 1m precision but if it happened to be exactly on the wrong corner of some OS grid squares the smallest grid square possible could well be 'SD'.
They're harder, the record isn't actually inaccurate, its just the mapping to the UK's coordinate system that is causing issues.
Currently I've not decided what to do with these. Three options I've thought of 1) Reject them, 2) Accept them (the underlying data specifies that it is Lat, Long and accuracy so the data is scientifically useful), 3) Edit them to put in a more plausible grid reference (accepting that might introduce a small inaccuracy in the actual location). Each of these has problems so for now I'm just leaving them pending.