My standard ‘circle of confusion’ is 2 km around a fixed spot that is not my home. The vast majority of my observations fall within that circle. I will not use a more precise position because my personal safety is more important to me than anything else.
I am reluctant to leave a trail on iNat or any other platform. Where I go, and when, is my business, not anyone else’s.
I do not trust/like Google. I have location services and any other Google stuff disabled on my Android phone.
Nevertheless, in a few instances where the exact spot of an observation mattered to a researcher, I provided the exact coordinates, by email, with a detailed description of the specific place (vegetation, habitat, roads, human presence, weather, whatever was relevant).
I use a trail GPS, which can find satellites from just about any location, and log the lat-long for each obs in the field. And I check the coords online before posting. It’s time-intensive and not recommended for those who post large numbers of obs. Just my current workflow.
I think it was @dianestuder who said we bear the responsibility for the IDs we post, and I suppose that goes for geolocation data as well.
I recently posted a Sonoran scrub oak which was growing in a rather desolate area, probably near the edge of its habitat tolerance. After posting I noticed that another observer had logged the same oak five years earlier. Interesting to see that the tree had changed very little. Couldn’t have been certain it was the same tree without accurate location data.
But iNat users vary widely in their reasons for posting and the technologies they use. I wouldn’t want to exclude observations solely due to large location radii. But there are good reasons to be accurate.
My locations on iNat are set by hand, I don’t have GPS. Anyway don’t have coverage on our mountains off the well beaten track. Wetland or forest - I would try to include that info either in a wider habitat view, with water or other wet preferring species, or else in a note. That info is useful to me as I learn new species, or tease out which species is this one. Doesn’t help that the location ‘on the trail’ is a circle - 95% of the area is off the trail and I was not!
that’s fair, but I’d basically be running tracking apps ALL THE TIME then. I don’t really go out with the intent to observe things for inat. I’m going out to do other stuff and I take pictures of things I encounter that interest me. some of it I’ll post to inat. for some of the things I do, I’m already using a Garmin fitness GPS to track my activity, but I’m very much not interested in the old school process of geotagging my photos using timestamps. I’ll draw a large-ish accuracy circle and call it done.
hm, I didn’t know such a setting existed. I just changed it. I still don’t care for the multi-step workflow to open inat then start an obs, then take the picture using inat’s camera app. a lot of the stuff I take pictures of is moving and fleeting, so stopping, pulling out the phone, unlocking the screen, opening the camera app, taking the picture is still not quick enough. but it’s simply not possible to get faster. well, I could disable the lock screen, but I have apps on my phone for work that require additional security settings.
what I could possibly live with is using the photo app to grab the image with the fewest steps I can manage and THEN creating the observation, updating the location from the inat app with less rush. It’s more or less the process I use with audio observations I obtain from the BirdNET app, which will export the audio file to inat, but not ANY of the metadata.
The line between upland and wetland is fluid, and is seldom discernible from an aerial image. There are facultative upland species that in some areas would be facultative wetland.
You’d be surprised. This is a lot of my job and I can get it right most of the time. Though it’s true that the boundary often isn’t well defined especially if you care more about ecology than legal boundaries used for wetland delineation.
Users who are concerned about broadcasting their location while in the field can wait and upload the observations after they’ve left, perhaps days later. There might be some conditions where even that is too revealing, but it might help for some.
This is a good idea for places you’ve only visited once or only visit sporadically, but if it’s something indicative of a routine (observations made at home, work, or between the two places, observations made somewhere you regularly hike or jog, etc.) it might not be ideal.
I usually use the “obscure” feature when I don’t want to divulge the exact location or use a 500m-2km radius. There are just certain locations I don’t want to divulge and I understand other people might not be comfortable with broadcasting that either. For example, your house, your workplace, etc. are things you could conceivably want to keep private.
I do enable sharing for things like projects and I’ve made a note on my profile that I’m happy to share those locations for situations like scientific research.
There have also been situations where I don’t remember the exact spot, just the general area. People have made some good points about limitations. Sometimes phones just aren’t powerful enough to have a ton of apps running in the background without causing major lag or battery drain, some locations don’t have a strong wifi signal so people are just uploading their observations later, or any number of issues. I agree that people should aim to be as precise as possible when they can, but that’s not always possible. I’d also say that in many cases it just doesn’t matter because they’re species present in a whole area.
I regularly review observations in the US state of Alaska where the state name is used in this manner. The accuracy zone includes both the Arctic Ocean north of the state, to a huge area of Pacific Ocean west to include the Aleutian Islands, to down south inclusive of much of British Columbia, Canada. I’m not sure it’s a “problem” per se, but it would be great to have language or something that results in the OP’s title: a more precise location for these observations. My sense is that 1 in 10-20 respond to a comment suggestion to revise the location. Many are users without a long site history, and I haven’t encountered anyone yet that responded with an intent to use such a large accuracy circle.
As mentioned above, sometimes it’s simply the GPS on the phone/app/camera that’s not accurate – I’ve found that when I toss a message to the user, it’s remedied with a pleasant conversation. :)
I usually upload observations from my DSLR that doesn’t have GPS. I upload in bulk and often can’t remember EXACTLY where each observation occurred, so I use a pretty generous radius circle and I’m not very worried if a few of my observations were outside of that radius a little. Since most of my observations are birds, I don’t think exact location is very important. Plant observations are different, and I had one observer ask for a more precise location of one of my observations since the plant was rare for that area. I was happy to provide it. My observation locations have a radius of about 100m. Is there any reason for the data to be more precise than that? And if so, perhaps iNaturalist is not the place to expect that level of accuracy in the data. It seems to me that forcing people to be extremely accurate with their observations would discourage people from using the app at all. iNaturalist is crowd source data and it has its strength and its weaknesses. I think discouraging uploads for the sake of high fidelity GPS marking would be missing the point of iNaturalist: to encourage a great awareness of the natural world, to keep a record of your observations to share with other people, and, perhaps most importantly for the future of this wonderful app, to enjoy yourself.
I am one of those who didn’t know how the precision field was filled in. Most of my iNat records are from geotagged camera pictures, not from my phone, and those went in without the precision field being filled out. I only recently discovered this, and now I manually enter an appropriate precision as I put in observations. But now what do I do with the several thousand records with no precision? Is there a way to batch edit them? I normally run with a 20m precision, unless I have reason to enter a larger or smaller one. I also noticed that photos taken with my iPhone have a precision entered automatically, some of which are unrealistically small. A 2 m precision is pretty hard to get down in a canyon. Does anyone know of a way to batch edit the older records, or will I need to go back and manually edit each one individually? Oh what a pain that will be!
Never mind, some imprecision in places like that is understandable.
I referred to other observations. For example, today a handful of observations (of cultivated plants…) centered on a town in the middle of Italy with a radius ranging from 6 to 10 km.
For my own work, I include the positional accuracy in the data download, and then I filter out observations with too wide a positional accuracy during my processing. It appears that the download page has now been modified so that the filtering can be done as part of the download.
If you need to filter earlier in the process (to ignore these observations during the ID process), I guess you could try via the URL as suggested in another post. My accuracy threshold is much higher than yours, and I only find a few observations exceed it, so I don’t worry about filtering them out from my ID process.
It’s possible to edit (add or modify) the “Accuracy:” information for a batch of observations, on iNaturalist’s website. Go to ‘Edit observations’ page; select the observations; click ‘Batch Edit’; unfold ‘Batch Operations’; click under ‘Acc (m)’ to reveal the hidden (?!) textbox; enter the desired “accuracy” (whatever is implied by this); click ‘Apply’; click ‘Save all’ at the bottom of the page.
Thanks for the validation, @ben56. I also have been uploading my observations from my DSLR, as well as a camera that has GPS (but I haven’t figured it out). I do not recall anything in the Help page regarding the need for precise location data, and my default for location accuracy is ‘obscured’. I do that for privacy, as most of my observations are around my home, sometimes to protect plants vulnerable to poaching, and because I honestly didn’t think it was particularly important.
Conversation about this topic makes me want to take my toys and go somewhere else.
Good point. I’ve had the same problem doing a quadrate, ie 100 m square, to survey a 10m wide creek. I don’t know the answer, other than to argue with the boss.
Maybe in ‘notes’ or ‘comments’ write that it was on the road. I’ve done similar with a group of trees growing along a linear feature. Basically I wrote a paragraph describing it.
Hmm…
this thread started out with:
How to encourage users to provide a more precise position
and has ended up with:
These are the problems I have and how I deal with them.
re: the first bit, my comment is:
I agree with the general thrust of all the above. Better precision would be good.
But law of unintended consequences. People who get told that 2km, or 100m, is not good enough will try to get their precious observations accepted by claiming much greater precision. So fish on a named mountain top, precision 5 meter. Something next to a named homestead, precision 10m, when it couldn’t possibly have grown within 10km.
I’ve written in other threads, when I’m working with a database produced by iNat + other sources, I’m happy to use or exclude obs with based on a stated location uncertainty. But I’m always fearful of all the ones (and I’ve seen it done), that wrongly claim great precision, just to get their observation noticed.