Incorrect border for cities in Silesia (Poland)

Bug Reports
1

Hello,
I found in my calendar view that some of observations were matched to wrong city. For example my walk through center of Katowice is partly attached to Siemianowice Śląskie:
https://www.inaturalist.org/calendar/wwwojtus/2025/4/8
Similar thing with Tychy:
https://www.inaturalist.org/calendar/wwwojtus/2025/4/14
There is listed place Katowice City but real Katowice are far away from my route that day.

Then I found that there exists correct places for Tychy and Katowice:
https://www.inaturalist.org/places/tychy
https://www.inaturalist.org/places/katowice
but automatic matching is done for incorrect:
https://www.inaturalist.org/places/tychy-city
https://www.inaturalist.org/places/katowice-city

I also found that there are two places for two other cities from Silesia, but none of them have correct boundaries - those for Świętochłowice:
https://www.inaturalist.org/places/swietochlowice
https://www.inaturalist.org/places/swietochlowice-pl-sl-pol
and those for Siemanowice Śląskie:
https://www.inaturalist.org/places/siemianowice-slaskie
https://www.inaturalist.org/places/siemianowice-slaskie-pl-sl-pol

I didn’t check other places from Silesia Metropolis (https://en.wikipedia.org/wiki/Metropolis_GZM) - maybe there are more such issues.

My question is why there are such duplicated places and what version of GADM maps are used?
It seems that now there are correct borders which I’ve checked here: https://www.geoboundaries.org/visualize.html?country=POL&mainSource=geoBoundaries%20(Open)&comparisonSource=GADM%20v4.0.4&mainLevel=3&comparisonLevel=3

Shouldn’t those place be merged (each city with one palce) and boundaries updated where needed?

I set a flag for one place and from discussion there it seems that there is no easy solution:
https://www.inaturalist.org/flags/747946

2

Most of the administrative borders on iNat come from GADM. GADM has tons of problems and is not very accurate, especially for sub-country-level borders. There have been many discussions about this in the past. Unfortunately, it is the best freely available border dataset as far as quality and comprehensiveness. But it also hasn’t been updated since 2022. We’ll probably have to move to something else eventually. It’s honestly kind of ridiculous that there are no good free border datasets available. If you’d like to explore some of the options out there, take a look at https://www.geoboundaries.org/visualize.html.

1 Like
3

Ok, but latest GADM map seems to have correct borders for those places - it is visible in the comparison I linked above.
Comparison is between GADM and geoBoundaries and differences are minor.

I think that geoBoundaries are correct, at least at places I was checking.

4

I don’t know when the last GADM update for iNat was but it was years ago. See:
https://forum.inaturalist.org/t/current-gadm-version-on-inat/39340

5

See https://github.com/inaturalist/inaturalist/issues/3522 which has been untouched for some time.

1 Like
6

So to sum up - we know that there are many problems with places, we know that GADM maps in iNat are old and we know that newer maps have corrected issues.
Maybe it is finally time to update? What can go wrong?
I’ve got a good feeling about this :-)