For example, consider a golf course. In England it is common for each golf hole to be separates by areas of trees. These can be single lines of trees are quite large groups of 50 to 100 trees. These are not really woods - at least not as I understand them. They are usually manages quite carefully so maybe they could be a forest. However ten trees doesn't really seem like any sort of definition of a forest.

asked 04 Dec '13, 22:14

RobChafer's gravatar image

RobChafer
220111117
accept rate: 0%


My practise for mapping (including on golf courses) is for any area of trees is to use natural=wood as this simply defines the area to have some trees in it! Although sometimes I don't map all the 'smaller' trees - it's a judgement call on to how keen you are / how much time you have to map it.

For reference you can view the OS Streetview Layer which contains wooded green areas for even smallest groups of trees.

NB There is the option for landuse=forest too (see http://help.openstreetmap.org/questions/324/when-should-we-use-landuseforest-rather-than-naturalwood for further discussions).

In the absence of any specific name on the group of trees probably means it's not a wood in the sense that I think you've thinking of: i.e. a reasonably largish area sustaining typical woodland flora and fauna.

For golf courses, especially for big single tree(s) that one considers to be part of the hole or an obstacle on it, I like to mark these individually with a natural=tree at the appropriate places.

permanent link

answered 04 Dec '13, 23:27

robbieonsea's gravatar image

robbieonsea
90411325
accept rate: 16%

+1 for natural=wood and natural=tree. You can also use natural=tree_row and barrier/natural=hedge in some cases.

landuse=forest should IMHO be reserved for forestry / harvested wood.

(05 Dec '13, 11:54) Vincent de P... ♦

Yes, I forgot to mention natural=tree_row.

However that tag languishes in limbo land. As far as I'm aware no major renderer draws it and no editor supports it as a default preset option.

Hence there's little incentive for a mapper to map it this way if they are going to map it at all. If one is keen enough then you'll probably go for mapping each individual tree.

(05 Dec '13, 21:25) robbieonsea
2

The osm fr style renders natural=tree_row, along with many other niceties.

(06 Dec '13, 13:06) Vincent de P... ♦
2

@Vincent This style is really pretty.

(06 Dec '13, 14:16) scai ♦

Despite robbieonsea's answer which probably represents current practice, other than for avenues of trees with tree_row,there is no good way of doing this.

The example of the practice of the Ordnance Survey in their StreetView tiles is not really one to follow: many green areas on these maps are single large trees in grassy parkland, low scrub and at least one case a wetland with no trees at all: sedge bed mapped as woodland

This demonstrates why overloading natural=wood or landuse=forest is not particularly helpful: it solves the problem of rendering, but reduces the value of the data. Consider someone wanting to use OSM to evaluate wooded landcover they would need some means of excluding all the little patches of woodland on golf courses and in parkland. Most national forest agencies, such as the Forestry Commission in Britain have detailed rules as to what counts as woodland. These are a useful guide for our own mapping in OSM.

You have the following alternatives:

  • Follow robbieonsea's advice and use natural=wood or landuse=forest, with the longer term disadvantages outlined above.
  • Use natural=tree with a closed way (area) or open way (line). This has been suggested for a long time as a way of mapping such groups of trees. Although there are 2000 instances of this on OSM (taginfo, OverpassTurbo query for Germany), some are nothing more than oddly octagonal individual trees (not recommended), so again the meaning of the tag is unlikely to be clear.
  • The denotation tag has a value which is meant for identifying trees in groups, denotation=cluster. Unfortunately a mass edit added this particular tag to well over 300,000 tree nodes. Again the (abuse) of the tag has devalued its meaning.
  • Use method 1 and add another tag to disambiguate these groups of trees from true woodland.
  • Invent some other tag entirely.

Personally, I think adding an extra tag is the best that can be done (penultimate suggestion). You get the benefit of existing rendering rules, but enable these areas to be picked up or excluded in queries. The most obvious, but somewhat clunky ones are group_of_trees=yes and tree_group=yes (by analogy with row_of_trees and tree_row).

permanent link

answered 08 Dec '13, 11:41

SK53's gravatar image

SK53 ♦
22.3k46228350
accept rate: 20%

I can understand why one would want to distinguish two things that happily shared natural=wood before, but the separation is rather unclear. Before discussing about tagging, we should get a clear definition of which is which, how to recognize them, and why they are useful to separate.

(14 Dec '13, 09:01) Vincent de P... ♦
Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Question tags:

×32
×17
×14
×12

question asked: 04 Dec '13, 22:14

question was seen: 5,335 times

last updated: 14 Dec '13, 09:01

powered by OSQA