[Imports] Polish Forests - Imports

andrzej zaborowski balrogg at gmail.com
Wed Oct 30 23:38:24 UTC 2013


On 30 October 2013 23:32, Paul Norman <penorman at mac.com> wrote:
>> From: Grzegorz Sapijaszko [mailto:grzegorz at sapijaszko.net]
>> Sent: Wednesday, April 17, 2013 1:41 PM
>> To: imports at openstreetmap.org; talk-pl at openstreetmap.org;
>> zarzad at openstreetmap.pl
>> Subject: [Imports] Polish Forests - Imports
>>
>> to: imports@, talk-pl@, zarzad@
>> Dear all,
>>
>> The data which is considered to be imported contains the borders of
>> forest's sections (small areas of the forest, about 5-20 ha) divided by
>> forest's headquarters.
>
> After being pointed to changeset 17263910 I think this question needs re-examination, now that we've seen the data.
>
> For an overview of the data, see
> http://pnorman.dev.openstreetmap.org/imports/review/polish_forests/forest_blocks.png.
> It seems to be typical of the data this import consists of. I've
> highlighted a few problems
>
> - There are several areas which are not forests

Yes, the data being imported is actually limits of parcels belonging
to the Forestry Service, which someone on the forums estimated
corresponds to forest boundaries in a grade of 80%.  It was argued
that it should be instead imported (if at all) with boundary=* tags --
there's one which is already established for that purpose (iirc mostly
in Russia).

> - There is a section which is split into many small polygons, when the
>   entire area is forest.
> - There is an a regular grid pattern which doesn't seem to correspond
>   to anything on the ground.
>
> Unfortunately, this import didn't get much review on imports@ at the
> time.
>
> Are the divisions between these squares observable on the ground? If
> not, I think there's a data modelling error here and all the squares
> should be merged into one forest, as they are indistinguishable on the
> ground.

I thought so too, but the internal divisions are marked by posts on
the corners and it is argued that locals, and mushroom pickers, tend
to use the numbers of these parcels.

>
>> Additional information about the data is located under >
> http://wiki.openstreetmap.org/wiki/Pl:Importy/Lasy
>
> There's also a couple of documentation deficiencies
>
> Firstly, could one of the importers fill out http://wiki.osm.org/wiki/Key:adr_les to document that key. Why does it have "   -    -" in it?

The data provided by the forestry service had this in a shapefile
column with a constant length, this seems to be the main reason.

A bigger concern though were some reports on the forums of importers
replacing first-hand surveyed landuses traced from imagery with
incorrect data.

Best regards



More information about the Imports mailing list