Improving Location Accuracy on observations

On my new smartphone I tried that scheme with going into plane mode and turning GPS on, it works almost perfectly. Though it uses double GPS and I can’t say how good it will be with an ordinary system, still it’s much better than standing next to each object waiting for app to work.


I was wrong to assume! It’s always interesting to understand more about the different ways that people use the platform. Happy observing :)


That’s true. I think in terms of more general mapping, watching range changes, etc. For associating plants with microhabitats, more precise data are needed.


Did you get one of those new Huaweis with the dual-band chip? I’ve been wondering about those for when my current smartphone dwindles down to just a few hours per charge.

No, it’s Xiaomi mi9, so kinda the same thing.

The discussion here seems to be largely with regards to location accuracy relative to a the use of a smartphone. Pardon me if I go a little off that thread but I am staying within the overall subject of the thread as stated in the topic.

Most seem to be focused on the accuracy of the handheld recording device and not the accuracy of the location of the subject. If I am standing 25 meters from a tree to get a picture of it and my location accuracy of my device is 100% accurate to the centimeter (great precision) , would not my best accuracy of the subject be 50 meters? This is taking into account that I do not know where I am relative to that tree (north? south? east? you get it)

I will say up front here that I will submit things that have a very large (in)accuracy. If I could do things perfect I would but some times it is the best I can do given the tools and situation I have.

Most of my observations are of fauna. I have a 600mm lens on a DSLR that I use (info that is found in my automatic metada). I will normally try to crop my pictures of my subjects so that it is easier for identifiers to find the subject. It is not uncommon for me to have a subject that could be 250 meters away and may be only identifiable in a painterly way (a little fuzzy). I usually do not know which direction that subject was in and if it was a shot of an raptor in flight I have no landmarks. If I know exactly where I am, my accuracy is at best 500 meters. If my location device has a large (in)accuracy, my subject location becomes even more rough. (Please correct me if I am wrong).

I was also recently travelling several places where I had only a rough idea of where I was with either no or spotty cell phone reception and near the equator or the southern hemisphere where it seemed the GPS was somewhat spotty because of satellite availability (Again, please correct me if I am wrong).

Because I am observing fauna, and if instead I was using my phone app for this, I would not often get more chances to take repeated shots to improve accuracy (however I could take a picture of my foot).

I get why accuracy is important and I am all for helping to add data but I think that understanding peoples’ limitations and why they are here can go a long way to help this dialogue of improving location accuracy.

By the way, I used a tool here to figure out how far away my subject could be.


That’s something I’ve been wondering - is the location “where was I when I saw the organism” or “where was the organism”? Because the second one is much harder to give with any precision indeed :)


It’s where the organism was.


In situations like this, such as taking a photo of a cetacean from a moving vessel at sea, I do the same as you - I manually change the accuracy (precision) value to encompass where the organism was. That can result in a large value, even 500 or 1000 metres. It’s not a perfect solution (because we’re losing the information that the GPS fix was very precise) but is the best way I’ve thought of to deal with this. In most cases like this, animals are large and mobile and 500 metres is a trivial distance for them to travel.


I started a topic asking the same question a few months ago, and the general consensus seems to be where the organism was. See:

1 Like

Absolutely. My observation of a humpback whale here is from the end of a pier so I know where I was, but, it illustrates that a subject could be quite far away and still be identifiable. In this case, I have calculated a distance of about 1.5km away. This calculation is based on an estimate of how large the subject or part of the subject was. I think it could have been at least 2km away and still be identifiable at least in series as I submitted.

If I had been at sea with this sighting and not noted the direction I was facing then the accuracy with a precise GPS location would have been 3km at best. I personally would also want my accuracy ring to enclose the subject and not have it at the edge of that ring. Also note that the subject is mobile during this observation - is a subject moving laterally or moving towards or away from my position?


In theory yes but in practice if someone else went to that same point they would presumably be able to see that tree too. True that data won’t work for super high resolution analysis like trying to determine exact aspect but otherwise it does the job. That being said when I photograph a plant from a distance I do try to edit and move the point to where the plant actually is if I know where it was.


Hey guys,
it seems like a really interesting topic and I am sorry I didn’t had time to read all the replies.
For me higher accuracy is always better and it is something I was forced to think about last few days. I made quite a few observations in last few months and I realized some of them are misplaced…a lot. First I blamed my phone (that was not cheap at all!) and than I made a lot of research and realized a lot of smarphones have the same issue. Sometimes they just misplace the photo geo tag.
I also use navigation app called OruxMaps and while I’m doing a tracklog, I can make perfectly accurate points, in that app. My photo geotags and also my iNaturalist observations at the same time are innacurate as usual. I remembered I used app for mapping our local flora that worked in a similar way as iNaturalist, only it would start a tracklog everytime when it was activated. That way the points were more accurate. I suggest iNaturalist developers they think of starting a tracklog as a background process in a new versions of the app in order to get more precise data.
What are your thoughts?


I do the same, use a separate app to continuously log my GPS while iNatting. Incorporating it into the app has been suggested by a few people including myself :)


How inaccurate are the dots? If they’re a hundred yards or so, it may be that iNaturalist and your phone are using a different map datum.

My phone (and my son’s phone, both Android) has been acting badly for the last 3 weeks. I had it set on “High Accuracy” which means it will use wifi hotspots and cell towers in addition to GPS. It’s been useful for navigating big box stores but it seems to be giving wildly inaccurate positions lately. Using iNat, eBird, and Trackker, my position has bounced all over the neighborhood. The furthest are 1.5 miles away. My eBird tracklog had be cover 23 miles while sitting still in the backyard for 90 minutes.

Setting it to “GPS only” solves the problem. Weird, it’s never behaved this badly in the past.

no, it’s just that phones won’t constantly look for your current location, unless you’re running something that forces the phone to do so. (constantly looking for location can eat up a lot of power, especially on older phones.) when you take a photo in your phone app, it will add coordinates to the photo that represent the last location that the phone picked up. so if you’ve moved significantly since then, then the location added to your photo metadata could be significantly off.

in my opinion, the easiest thing to do is to open up your map app, find your current location, then switch to your photo or iNat app to take a photo.

if you can use GPS, it will generally be more accurate than wifi and cell tower location.


Yes, the “high accuracy” settings on cell phones are very misleading. Keeping cell towers and wifi completely out of the picture, by using airplane mode, will give you the best accuracy, especially if you also run a separate mapping or GPS app in the background to keep the GPS active all the time. The battery used to keep the GPS active will be offset by the battery savings of being in airplane mode.

1 Like

I feel like the application of the term “high accuracy” to those settings should be grounds for a false advertising suit. Once in a while when I use google maps to navigate a drive it will set my location settings back to that without my realizing it, and then when I go out and record a track in some remote location a whole day’s worth of gps logging will just switch back and forth between two cell towers from opposite directions, each a good dozen miles from my actual location. Performs poorly in town too. It’s like it should still be some kind of beta feature but instead it’s the default setting!

1 Like

I just loaded OruxMaps and I can get it to create a GPX file on saving a track. I can also email it to me as an attachment so I can reference it for putting GPS data on my photos. Is there a way to have it automatically email it to me?