Making your observations more valuable for conservation - location uncertainty

We've been busy reviewing records in the project and have to say we're blown away by the number of people submitting so many observations that can be used to update the provincial record - thanks to all of you for your hard work!

For now, we wanted to share some tips for improving on two common issues we have encountered when reviewing records: location uncertainty.

Location uncertainty is really important because it tells us how accurate the reported location is. Most handheld GPS units (including phones) can record locations up to about 10 m accuracy. The more accurate the location, the better our ability to ensure the correct location is identified as being important to a species.

Issue 1: no location uncertainty reported

You can create a record in iNaturalist and record the location perfectly, but in some cases there will be no location uncertainty attached to it - just a location and coordinates. In these cases we don't know how accurate/inaccurate the reported location is which makes it hard to assess the record, and especially whether the location is important for conservation (e.g. is the species in suitable habitat or likely just passing through an area). Some times, phones can't get a GPS signal and use triangulation from cell towers to estimate a crude location (or just report the location of the nearest cell tower). The most common cause is taking a photo and then importing it into iNaturalist.

When you take a photo with a camera or phone that geotags the photo, it records the location, but not the location uncertainty. So whenever possible use the iNaturalist app to take the photo, or, bring the photo into iNaturalist right away, and then manually use the app to calculate the location (and location uncertainty). If you're someone who takes photos with a camera that geotags the images and then later imports them to iNaturalist, consider using the app in the field to create the record (without the photo) and then import the photos later when you're on a computer. Alternatively, review your records and manually add in the location uncertainty.

If you'd like to see your records in the NHIC Rare Species of Ontario project that are missing location uncertainty, use this url and replace "username" with your username. You can manually edit these records to add a location uncertainty value if you can estimate it on the computer map.

https://inaturalist.ca/observations?acc=false&place_id=6883&project_id=nhic-rare-species-of-ontario&subview=table&user_id=username

Issue 2: very low location uncertainty reported

Very low location uncertainty (depends on the species/location but generally anything over 1 km) records are difficult for us to use because they are hard to evaluate their importance for conservation and also because they may "swamp" records of higher precision.

In some records we've come across, the uncertainty distance is very large (kilometres or even tens/hundreds of kilometres). This could be the result of bad GPS signal or you genuinely don't know exactly where the record is from (common with older records). Sometimes this is done on purpose by a user who thinks they are protecting the location by reporting a made up location with a big uncertainty distance.

In the first case, if you can refine the location by looking at a map - great, please do so. Usually when we've asked people they are able to estimate the location when looking at maps (imagery and trail maps are very helpful for this!). Please don't guess at this though - the location uncertainty should reflect the area you are sure about.

In the latter case, there is actually a really great feature of iNaturalist that lets you obscure the location of a record, while still maintaining the exact location. This is called geoprivacy. When you set the geoprivacy of a record to anything other than the default (open), only you, people you trust, and curators of projects you have given permission to (like this project) can see the exact location; everyone else sees a randomized point somewhere within the same ~20x20 km square as the actual observations (for obscured records) or nothing at all (for private records). Please note that for most rare species in Ontario, the geoprivacy is already set for the species as a default.

We'd encourage everyone to review your records with low location uncertainty and edit them if you can. Here's a link you can use to search for your records with uncertainty greater than 500 m (be sure to replace "username" with your username).

https://inaturalist.ca/observations?acc_above=500&place_id=6883&project_id=nhic-rare-species-of-ontario&subview=table&user_id=username

Thanks everyone, keep those rare species reports coming!

The NHIC team

Posted on October 25, 2019 03:39 PM by mikeburrell mikeburrell

Comments

I have a default location filter on my inaturalist map, and when I followed this link the only posts I saw were those inside the filter. I had to delete this filter to see all of my NHIC listed items missing accuracy information (I don't use the app so that was all of them). Thanks.

Posted by ajpto over 4 years ago

This is quite a coincidence. Just today we were going through records of our Help The Turtle project and found the same problem. Just over 50% of the observations had coordinates of >50m accuracy or no accuracy at all. Looking at different, not-at-risk species on iNat.ca, it was a very similar situation. I sent a notice to folks at iNat.org and the global iNaturalist network leads to see if we can come up with a platform-wide solution.
And looking at that link, I'm guilty of a few of those observations myself :-( I'll fix those.

Posted by jpage_cwf over 4 years ago

Thanks @ajpto - I've updated the urls so they should over-ride your default location. Let me know if it doesn't work.

Posted by mikeburrell over 4 years ago

Thanks @mikeburrell. That works to override the location filter. Looking at the comment from Help the Turtles I'm wondering whether I should be fixing all my entries, which leads to the question of whether there is a method for bulk editing of the accuracy field for people like me who upload from cameras to the website and have mulitiple past entries with accurate GPS coordinates and blank accuracy fields (unless of course I'm the only one who didn't know this field should be filled in...).

Posted by ajpto over 4 years ago

You can bulk edit the accuracy but I have to say I wouldn't recommend it, because there is a pretty good chance you'll just be making something up which could be wrong, and wrong data is worse than missing data.

But, if you want to bulk edit records, you can do so by clicking "edit observations" under your user image (top right of the screen) when on the website. On the new screen you can then use the search function to find the records you want, and then "select all" on them, then use the "batch edit" feature. I must caution you again though that this should be done VERY carefully as you could ruin a lot of records very quickly.

Posted by mikeburrell over 4 years ago

Fully understood. I'll use my powers sparingly and only for good. Many thanks.

Posted by ajpto over 4 years ago

Good tips guys. However, suppose I add or improve location accuracy to a bunch of species in the NHIC projects. How will NHIC know that my location accuracies have been edited and/or improved?

Posted by burke_korol over 4 years ago

Hi Burke, each time we download the data we'll look for changes such as this.

Posted by mikeburrell over 4 years ago

OK, good, because I've improved a bunch of my vague accuracies. Also, I'm a little worried about hackers getting into databases of sensitive species. If they ransomware hospitals, why wouldn't wildlife traffickers attack NHIC or iNat? What data security do these databases have?

Posted by burke_korol over 4 years ago

Hi Burke, I'm not the person who can answer a technical question of that nature. I'd suggest you send an email to NHICrequests@ontario.ca if you want more than I can give you. What I do know is that rare species information is considered at a pretty highly level of sensitivity within the government so additional safeguards are taken when it comes to its distribution and protection.

Posted by mikeburrell over 4 years ago

I have a question slightly related to NHIC. I found this post on my phone the first time using "Activity", then the "News" tab. But for the life of me, I could not find it on the web site where it is much easier to try out the links you provide and type comments. I've book marked it now, but perhaps it would be worthwhile to use a similar infrastructure in both the apps and the web page. This post, for example, is quite crucial to your project and to many others, no doubt. But maybe only a few people see it.

In fact, I did a search for location accuracy from the home page and got nada.

Thanks,
Rob

Posted by robinlanark almost 4 years ago

Back on topic, I notice that observations I make with my phone app (because that's where most of my photos are) have a reasonable location accuracy by default. But I have not idea if I've set that. Is it because of the amount of zoom I had on my location map when I clicked the checkmark on the map maybe?

Posted by robinlanark almost 4 years ago

HI Rob,

Regarding location accuracy. When you use the phone app to record an observation it used your phone's GPS to record the location. At the same time it uses the GPS and recording the location accuracy at the time of the reading (just like a handheld GPS unit would do). If you manually open up the location on an entry in the app and zoom in or out then it will manually adjust the accuracy depending on how zoomed in or not you are. But in most cases it's a record of the accuracy recorded by your phone's GPS at the time of the observation.

Posted by mikeburrell almost 4 years ago

Hello, I believe I have spotted Gar in my bay, there is about 4, plus a vast bass nesting ground. I am concerned, as there is so many sea doos, that have been raving around here.
If there was no fish I would not be concerned, but there is about 50 nests.

Posted by naturecove123 almost 3 years ago

Hi @naturecove123, that's great. Can you get photos of any? Adding observations to the Rare Species of Ontario project will help the observations to be noticed: https://www.inaturalist.org/projects/nhic-rare-species-of-ontario

Check out https://inaturalist.ca/pages/getting-started-inaturalist-canada for more tips on posting to iNaturlaist.

Posted by carrieseltzer almost 3 years ago

Thank you.

Posted by naturecove123 over 2 years ago

Add a Comment

Sign In or Sign Up to add comments