[OSM-talk] Landuse vs Vegetation vs Landcover proposed cleanup at wiki

[Martin Koppenhoefer]

>
> 2014-11-30 12:53 GMT+01:00 Richard Z. <ricoz.osm at gmail.com>:
> The problem is not only that forest can be mapped as either
> landuse,landcover
> or natural.
>
>
I'm also interested in this topic. I think it is not a problem to have
different keys, because there are different aspects of the same "object"
that deserve their proper description. E.g. for forests, there are
currently these "main" tags (at least)

* landuse=forest which IMHO describes land used as a forest. This could be
also a "void" area where trees have just been logged and new trees are
planted now. This feature can be arbitrarily split, as the semantics remain
the same: land used as forest.

* landcover=trees this describes an area where trees grow. It can be a
forest, but it doesn't have to be one, can also be much smaller (a forest
needs to have a certain size in order to be a forest, otherwise the forest
ecosystem cannot live). These can also be arbitrarily split.

* natural=wood (or maybe also natural=forest in the future): there are at
least 2 ways of interpretation, one is "unmanaged forest" (because
"natural" is read literally as "mother nature"=unmodified by humans) and
the other is: a forest entity (i.e. a "piece of forest with a name", maybe
not even trees growing everywhere, could have clearings inside with no
trees, but which would still belong to the "forest" somehow).






We must also go away from the - "there is either rock, vegetation or
> residential
> area" model.
>


this depends a lot on the scale and generalization you use in your mapping,
but clearly "vegetation" is not in the same category as is "residential"
(because a residential area can have vegetation).



>
> Furthermore our vegetation model - "there be either forest or gras" is
> woefully
> inadequate.
> We need vegetation layers (ground, shrub level, under canopy, canopy,
> emergents).
>


+1 (but 5 levels seem kind of an exxageration, still I would support the
introduction, but would not expect that we will get a lot of this data for
all levels in the next time).



> We need lots of fine tuning in geology as well.
>
> Instead we need the possibility to say
> * 70% landcover is sand (+ material where known)
> * 10% larger rocks (+ main rock type)
> - spatial extension of those areas may not be identical
> * ground level vegetation is gras, covering XX% area
> * shrub level is Ocotillo, with 100 plants per 100 sqm
> - again spatial extension of those areas may not be identical
>


+-0, this is in the end about generalization. I requires a lot of
standardization and adherance to these rules in order to get useable data.
It works in the scientific context, but I'd question if we can transfer
these concepts into osm.


cheers,
Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20141201/26e6cb43/attachment.html>