[Tagging] The endless debate about "landcover" as a top-level tag

Martin Koppenhoefer dieterdreist at gmail.com
Wed Jun 13 08:24:58 UTC 2018

2018-06-12 14:59 GMT+02:00 Paul Allen <pla16021 at gmail.com>:

> On Tue, Jun 12, 2018 at 1:14 PM, Martin Koppenhoefer <
> dieterdreist at gmail.com> wrote:
>> I didn't want to quibble and am seriously trying to understand you. To
>> me, a "group of trees" means a few trees, say starting from 3 to maybe 20
>> or maybe even 50 on the extreme end, usually something lower than 10.
> A group is an unspecified, and possibly indeterminate, number.  You're
> thinking of small groups. :)

yes, a group is generally an unspecified number, but context gives an
indication of the dimension. If you talk about a "group of trees", it is
IMHO a description of trees that are not in a forest, or are a very small
part of trees with specific characterstics in a forest. A whole forest will
not be referred to in natural language or have "forest" in the name (what
you were writing about).

While there is some uncertainty about the definition of what does
constitute a "forest", it is often clear for few trees that they are _not_.
A forest means more than trees, it means microclimate, soil structure,
habitat. There is a minimum size for an area with trees which is required
to form a microclimate, if there are too few, a forest will not form.

A group of trees is something like this:

This is not a group of trees:
it is a forest.

> Come up with a better term if you think it necessary.  I was trying to
> convey that this was a tag that applied to
> anything from a copse to a forest.

yes, the tag is applied to any kind of area with trees, but the word
"forest" is not.

> What we're trying to come up with a suitable name for is a tag to use
> when trees are too closely spaced *or trees which are not closely spaced
> but cover a very large area) that
> tagging individual trees would be far too much effort.

we actually do have a tag for this, we even have 3. ;-)

What we are lacking is a way to tell that a nature reserve is mainly about
a forest (e.g. a national forest), even if/although not all of the area is
covered by trees. Right now you can only see this from the name, if you
understand the language.

> ...  The tag landuse=forest is documented about marking areas where trees
> are grown
> to be logged, and which are periodically cut down,

Although after a lot of modifications, the wiki is less clear than ever,
there is no requirement of "logging" or periodically cutting down on the
relevant wiki pages. The tag definition page
https://wiki.openstreetmap.org/wiki/Tag:landuse%3Dforest states:
"Forest. Sometimes considered to have restricted meaning "Managed woodland
or woodland plantation". "
"*Managed woodland* or woodland plantation. There are major differences in
the way this tag & natural <https://wiki.openstreetmap.org/wiki/Key:natural>
=wood <https://wiki.openstreetmap.org/wiki/Tag:natural%3Dwood> are used by
some Openstreetmap users. Some use this tag for land primarily managed for
timber production, others uses if for woodland that is in some way
maintained by humans, some for any forest. This problem is explained in the
page *Forest <https://wiki.openstreetmap.org/wiki/Forest>*."

i.e. the wiki states the requirement for landuse=forest to be about a
forest, but links to another page https://wiki.openstreetmap.org/wiki/Forest
which reduces the requirements:
"A *forest* or *woodland* is an area covered by trees. Two different tags
are used to describe this: natural
<https://wiki.openstreetmap.org/wiki/Tag:natural%3Dwood> and landuse
<https://wiki.openstreetmap.org/wiki/Tag:landuse%3Dforest>. There are major
differences in the way these are used by some OpenStreetMap mappers.
Situation is complicated as different people advocate different,
conflicting tagging schemes. Depending on region there may or may not be
difference between areas tagged as natural
<https://wiki.openstreetmap.org/wiki/Tag:natural%3Dwood> and landuse
<https://wiki.openstreetmap.org/wiki/Tag:landuse%3Dforest>. Difference, if
any, depends on who mapped the area. As result nearly all data consumers
treat both natural <https://wiki.openstreetmap.org/wiki/Key:natural>=wood
<https://wiki.openstreetmap.org/wiki/Tag:natural%3Dwood> and landuse
<https://wiki.openstreetmap.org/wiki/Tag:landuse%3Dforest> as synonymous
tags for a forested area."

inside the forest there will usually/often also be areas where landcover is
> not trees, e.g. lakes, meadows, etc., but which are contained in the
> legally protected area. These areas should not be mixed up (IMHO) with
> landcover or landuse, as their boundaries are not depending on each other.

Now you appear to have diverted to a sidetrack of your sidetrack.

not from my point of view. This paragraph is about places like"sherwood
forest", which are about a forest area, but you shouldn't expect that every
squaremeter inside this area is actually covered with trees. It is about an
administratively defined nature reserve named after a forest.

> 1) Outer area (however we tag it) has name Sherwood Forest.

original landcover proposal thought to use "natural=forest / wood" for
these, but the key natural has since been amended with values like mud and
sand which smears the meaning of natural as describing a distinct object
(like a tree, a spring, a beach, etc.). At the current situation I would
either suggest protected_area or nature_reserve, or "place" to be a
suitable home for these toponyms referring to forests and woodlands.

> 2) Inside Sherwood Forest are one or more areas tagged landcover=trees and
> possibly lakes, streams and meadows.
> How we deal with that is (possibly) a different problem.

which problem? Simply tag lakes, streams and meadows.

I would exclude these from landcover=trees but include them in the place /
nature_reserve object naturally (unless they are excluded).

Right now, it's about deprecating landuse=forest because it's
> being used two different ways.

forget about "deprecating" landuse=forest, it is impossible, any better
system will have to be compatible (use different keys), or you'll end up in
an edit war. You can think about deprecation _after_ a new scheme has been
adapted by the community.

>   And we ended up on this massive diversion because I happened to mention
> that one
> reason people misuse landuse=forest because they encounter groups of trees
> that aren't being grown for logging in
> an area named (for example) Sherwood Forest.  The mention of Sherwood
> Forest was to underline why landuse=forest
> is a bad way of tagging things.  The fact that Sherwood Forest is a nature
> reserve is a side issue.

It is not a side issue that many things called "XYZ forest" are actually
nature reserves with different landuse inside. It is one of the main issues
in the question of forest tagging how to map these (where to apply the
name, how to state it is about a forest).

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20180613/2221148b/attachment.html>

More information about the Tagging mailing list