Why are the boundaries of my city drawn this way?

For a bit of context, Wrocław is a big city of around 800k. Forests are sparse here, and thanks to our multiple rivers there’s plenty of wetlands in and around the city. There’s also a big mountain range a bit further South.

I don’t quite understand why iNaturalist defines “Wrocław” as this specific fragment of the map.

It doesn’t make sense geographically - obviously what people normally call “Wrocław” includes the city centre lol. It also doesn’t make sense ecologically, because there is no clear reason to differentiate between its South-Eastern surroundings (included) as opposed to North-Western (excluded). They are practically the same.

So is there some valid reason to split Wrocław this way, or is this a bug?

Strange – the next level up for places, includes the whole city and also a swathe of countryside. Nothing in between.

But also other problems for the map in Poland
https://forum.inaturalist.org/t/incorrect-border-for-cities-in-silesia-poland/64908/4

This is the Powiat średzki https://pl.wikipedia.org/wiki/Powiat\_średzki\_(województwo_dolnośląskie) but the boundaries are from GADM and incorrect. It certainly should not include the city of Wrocław.

However, Powiat wrocławski is indeed just the SE of Wrocław https://pl.wikipedia.org/wiki/Powiat_wrocławski and it is probably what we see in the map in the OP.

The map of Powiat wrocławski on wikipedia shows a separate “powiat miasto Wrocław”:

And Wrocław is indeed miasto na prawach powiatu in the list of powiats https://pl.wikipedia.org/wiki/Lista_powiatów_w_Polsce and is separate from powiat wrocławski.

Disclaimer: I am not Polish, but can sort of understand.

Is this not because the “Places” boundaries are for that piece only? See the first two https://www.inaturalist.org/places/search?utf8=✓&q=Wrocław&commit=Search

Found the same problem with Nieuwoudtville - the town is not included in the map of the town at all.

But on the places subsection, the only entry for Nieuwoudtville is the Wildflower Reserve which is some ways out of the town https://www.inaturalist.org/places/search?utf8=✓&q=nieuwoudtville&commit=Search

A lot of places are user-created and there is a mismatch between places found by English spelling and transliteration of native names. In my own experience, for example, Moscow and gorod Moskva places differ by inclusion of two autonomous districts that used to be considered Oblast.

There are two duplicate iNat places with the same polygon. Both seem to be automatically generated since neither cites a user who created it, but as a curator I have the ability to edit one and not the other. Strange in multiple ways.

https://www.inaturalist.org/places/wroclaw-pl-ds-pol
https://www.inaturalist.org/places/wroclaw

I made a flag for the issue although I don’t know what we as curator can do about it: https://www.inaturalist.org/flags/847841

If you have a shapefile for the actual city that would help.

What’s the URL of the page you’re looking at?

Thank you, yes, this is correct. Let me try to untangle:


“Powiat” (“county”) is a small administrative unit.

“Gmina” (“provincia”) is a larger administrative unit that contains one or more powiats.


In case of Wrocław:

Wrocław = powiat miasto Wrocław. It’s a “miasto na prawach powiatu” - a city so large it constitutes a powiat by itself.

gmina Wrocław = powiat miasto Wrocław. It’s a gmina (larger unit) that simply happens to contain only one powiat (smaller unit).


The crucial part:

Powiat miasto Wrocław (“Wrocław City county”) and powiat wrocławski (…“Wrocławian county” ¯\_(ツ)_/¯) are two completely different and unrelated counties.

Powiat wrocławski is only named this way because it’s close to Wrocław. It could as well have been named anything else, just like other powiats surrounding the city are.

If you think this naming is weird and confusing as hell, it’s because yeah, it absolutely is. Still, it’s a pretty huge mistake to consider an area merely named after the city as the city itself.


My bug report and fix request:

In iNaturalist search, I can’t currently select the area that I’m interested in (which would be either Wrocław the city, powiat miasto Wrocław, or gmina Wrocław).

Both top results link to powiat wrocławski, which like I explained is unrelated and therefore completely useless.

The thread @DianaStuder linked suggests it’s a problem with a completely ancient and incorrect dataset iNat map is using. In that case, let me join people pleading for the update.

To explain my use case: I’m a brand new user, just learning to use the site. First thing I tried was searching for my own city, just to become immediately disappointed I seemingly can’t either subscribe to it or select it in search filters.

I’d be mildly ok with it if it was some small town, but it’s not. I would accept some workaround like “select gmina Wrocław instead”, yeah, it would be jank, but at least it would work. Currently I guess I could manually select the area on the map when I search for observations, but it does seem completely impossible to subscribe to it. It’s as if the entire large city vanished from iNat database :grimacing:

Search url: https://www.inaturalist.org/search?q=wrocław&source\[\]=places

2 top results urls, both actually covering powiat wrocławski:

https://www.inaturalist.org/places/wroclaw-pl-ds-pol

https://www.inaturalist.org/places/wroclaw

For Cape Town we have the default in the top right search box – which is Marine

So I have to filter to Cape Peninsula 2 (which is OK, since I know that)

For one particular obs, the dropdown at Details for Location, may, give you a useful location name


For Nieuwoudtville your next level is Namakwa !! or Calvinia ! neither of which is particularly useful for one town.

@senthe - Please explain what context your screenshot comes from. Most places on iNaturalist have several different boundaries that are used in different contexts, some of which are user created or imported and some of which come from external sources like GADM. There is not just a single definition of Wrocław on iNaturalist. User-created places can be edited, but GADM boundaries cannot. And if you are referring to the GADM boundaries, most municipal boundaries from GADM are wildly incorrect (especially outside of the US and Europe), and this problem has been know about for years. iNaturalist uses a very old version of the GADM data, but even the newest data isn’t that much better. @rademacherdan might be able to provide more information.

The context is search results.

What I mean by “definition of Wrocław” is the area linked to the “/places/wroclaw” URI.

Otherwise, yes, there are two links labelled as “Wrocław” in search results. They are both wrong in the same way.

Ok?.. I’m not sure if you read my reply above, but I explained in detail there that the issue is not about boundaries being a bit wonky or something. It’s about one area being swapped for another because they’re named similarly. A human manually trying to draw a boundary of a city wouldn’t make that mistake.

The problem is that the area is just not correctly labelled. The area that iNaturalist labels as “Wrocław” is actually powiat wrocławski. If it was labelled as “powiat wrocławski” then it would be completely fine to have it listed. It’s just a regular county. Unrelated to the city itself.

I have no idea who is responsible for mislabelling this piece of data. Or for the other piece of data, the city Wrocław, going missing (presumably as a consequence of that mislabelling?). It might be the outdated GDAM dataset or it might be iNat itself.

https://www.inaturalist.org/places/wroclaw

A similar issue appears for Białystok https://www.inaturalist.org/places/bialystok and perhaps other cities where the area makes a ring around the city but does no include it. However, in this case there is another place called Bialystok City available https://www.inaturalist.org/places/bialystok-city This kind of place for the city itself is missing for Wrocław.

@senthe we can create a new place for the actual city. Is there anywhere online that you can get a shapefile of it? Otherwise someone will have to create one, which will be rougher and less accurate.

if there aren’t official boundaries published by some government entity, you can always get boundaries from OpenStreetMap. see:

I guess it depends on the source of the map and the reason it was drawn up. Maps are imported to iNat from elsewhere, not created By iNat. Where I live in the county of Essex, England we have a terrible map that no-one living in the county would recognise. It is based on local council maps showing which authorities control which areas. The map we’ve ended up with shows the area controlled by Essex County Council, but it misses out three other areas, including the largest city in the county and one of the best wildlife sites.

I fought a long battle to get a proper map of our county, but it was never going to happen. Eventually someone (Tiwane?) suggested a workaround that might work for you. It involved making a project that incorporates separate maps for all of the areas you want. Good luck!

The Wildlife of Essex · iNaturalist https://share.google/8NNWNho2ANNZndtoq

Most names refer to pretty arbitrarily defined things. I like to use terms like “blob” to underline this, to bypass giving a false impression of a strict meaning.

“Adelaide” can refer to a 2km square, also known as the CBD (Central Business District), over half of which is houses.

“Adelaide” can refer to a 4km square, over half of which is parks.

“Adelaide” can refer to anything within 3km of the toilet seat. (Look at Google Earth)

“Adelaide” can refer to a capital city full of suburbs, 70 x 30km.

“Adelaide” can refer to everthing south of the last traffic light as you go north, so 400km north to south.

And my pet peeve is common names, many of which I have never heard in my life, or which I’ve more commonly heard for other things, but which iNat seems to want to carve in stone.

iNat, like lots of big bureaucracies, likes to give the impression of precision and meaningfulness.