From OpenStreetMap Wiki
Jump to navigation Jump to search

Diminishing woods

I've found quite many places, where a thin area (width of 5 to 50 meters) of wood has been left between for example some industrial lots or between residential areas that grew to almost touch each other from opposing directions. No one has presumably ever "maintained" that thin area, but by the definitions once just one dead or dying tree is removed, it ought to be changed to landuse=forest. Or a small patch of wood has been left within a farm field, where it is very doubtful that the trees have been maintained - only ones which have fallen onto the field have been removed.

And a larger wood that borders a residential area but where only those trees that are a potential threat to the houses, should they fall, are removed: to me it would seem silly to draw a landuse=forest of 25 meters width and outside of that an another area with natural=wood. Alv 09:56, 23 October 2008 (UTC)

This kind of woods/forrests are also very common in Sweden. They are typically not really "maintained", but rather "contained": Trees are removed if they might fall over nearby houses/roads/powerlines. Maybe we should use natural=wood, wood=contained (just a suggestion, feel free to suggest another way of tagging)? --SamuelLB 17:46, 31 October 2011 (UTC)
I checked with Swedish Wikipedia here and it says that "natural" wood means that is has been completely untouched for 150 years (the English wikipedia says "undisturbed" for 150-200 years). So I guess "managed maintained" should not be interpreted literally, but rather as "humans have touched it in any way in the past centuries". Which makes sense with the statement that most woodland in Europe is "managed maintained".

OSM uses different meanings compared to some dictionaries. The key 'natural' is on the OSmwiki defined as "Used to describes natural physical land features, including ones that have been modified by humans". With this definition it is fine to use natural=wood on 'maintained' areas of trees. It is not good to use the key 'landuse' unless you know that it meets the OSMwiki defintion of 'landuse' which is "Land use describes the human use of land" and for forests I would take that to mean some thing is taken, usually trees, for human use. Warin61 (talk) 05:39, 8 February 2019 (UTC)

Minimum age for a wood

In the densest areas around here, the military presumably cut down a large lookout zone for defence sometime around 1850. Naturally most of that has since then grown back and has only lately been conquered again for building, in slices. To me those remaining areas are "unmanaged unmaintained", but what's others' view on that? Alv 09:56, 23 October 2008 (UTC)

Density of woodland

Many areas that are not 'forests' have large areas of not so dense tree cover, often with housing or farmland mixed in - e.g. trees to shade animals. Maybe we should have a tag to indicate this sort of tree cover. Jamesks 8/01/09

Agreed. I suggest the following tagging natural=wood, wood:density=sparse. --Glebius 14:45, 16 December 2010 (UTC)
I note that the application of the natural=wood tag/value shows in the Simple tab as "Woodland" which is even more misleading as woodland is a type of low-density forest and in Australia it refers to discontiguous canopy cover as distinct from a forest. I'm disinclined to apply natural=wood to true woodlands if it implies a forest. I'm currently looking at two adjacent nature reserves, one which is dry sclerophyll forest predominantly populated by Eucalyptus polyanthemos, E. bridgesiana, E. rossii, E. mannifera and E. macrorhyncha … and the other which is woodland comprised of Eucalyptus melliodora and E. blakelyi, an endangered habitat and vastly different ecosystem to the aforementioned forest even though the canopy is comprised of trees from the same genus. What I'm saying is I'm not even sure that a wood:density sub-tag is adequate. Yes I could go into using the vegetation descriptors if I learned more about the taxonomy but at a high-level and from a rendering perspective forests should not be lumped in with woodlands. They look remarkably different, have different recreational uses, different ecological value etc --NathanaelB 03:54, 19 July 2013 (UTC)

"Woodlands are a broad category of vegetation in which stands of dominant trees are distinguished from those in forest and rainforest by their height, spacing and crown cover. Perhaps the most workable definition of woodland is that of Richard Hobbs: 'ecosystems that contain widely spaced trees with their crowns not touching'." - Lindenmayer, David, Mason Crane, and Damian Michael. Woodlands, a Disappearing Landscape. CSIRO, 2005. Print.

--NathanaelB 10:54, 20 July 2013 (UTC)
Originally, natural=wood was used to indicate just that "here are lots of trees", and the landuse=forest was added to that, when it was obvious that it was "used for forestry". The current usage is a result of an unfortunate mass edit years back, when somebody thought it was superfluous information; the words must be interpreted in the context of non-native English speakers trying to make sense of them (and trying to describe the usage), so dictionary and vegetation research definitions are mostly unusable guidelines for casual mappers, at least globally. Also note, that the "Woodland" word you refer to, is just one map editor preset label, and not to be considered in any way "official definition". But it gets the point across; "here you have lots of trees". Alv (talk) 07:59, 22 July 2013 (UTC)
'We' now have landcover=trees! Undocumented, but I assume this is "here be trees" of any type, use, age or density. It would be ause full tag where usage, type, and/or density detail is not known - say from imagery. With this tag in existence landuse=forest can be restricted to those areas where trees are used for some productive output (usually 'lumber' - wooden beams, planks etc). Then natural=wood can be used for areas of trees not meant to produce products? Warin61 (talk) 08:15, 31 January 2016 (UTC)


Changed from decideous to broad_leaved. Usually maps show if woods are coniferus or broad_leaved. For instance, Japanese Larch is coniferus and decideous, European Holly is broad_leafed and evergreen. Also added palm as a third distinction. Unsigned comment by User:Hawkeyes

Agreed, just evergreen/deciduous is really not appropriate (think, e.g., of tropical forests). Type categories like "broadleaf" and "coniferous" should be added, and maybe also "semi-deciduous". The different types could be combined, where appropriate: "type=broadleaf,semi-deciduous". --Ulf Mehlig 19:01, 3 May 2012 (BST)
Copied from discussion of landuse=forest:
Please avoid using type=* (for just about anything). A way with landuse=forest (or, rather, natural=wood) can have other tags which could (but shouldn't, either) also make use of type - and you'd have a conflict. The convention has always been to use descriptive keys wood=coniferous, and, say, leaves=evergreen/deciduous/... Anyway such drastic changes should go through the tagging mailing list at minimum. Current tagging is not like that. Alv 08:20, 3 June 2011 (BST)
So, maybe better foliage=broadleaf/coniferous, foliage_persistence=deciduous/semi-deciduous/evergreen? --Ulf Mehlig 19:22, 3 May 2012 (BST)
"We' now have leaf_type=needle etc and leaf_cycle=deciduous/evergreen etc. And perhaps in the future wood_hardness=soft/hard? Warin61 (talk) 08:19, 31 January 2016 (UTC)

Why not all woods?

Why can't this tag be used for all possible woods, even maintained ones (like natural=tree), similar to natural=water (that can be used for simming pools, even)? That the forest/trees/area hasn't been maintained for a long time seems like it should be a sub-tag. Alternatively I want natural=tree to be able to cover an area... :p No, landuse=forest doesn't cover all other options (and should maybe be orthogonal from use of natural=wood or natural=tree) /Mirar 19:51, 19 May 2012 (BST)

Because a few people pushed that definition a few years ago and now many mappers have learned the natural=wood vs. landuse=forest distinction as the "right" way to map this. It has always been a bit of a controversial issue, that's why the page has the three "approaches" listed under "Notes". Unfortunately, the first one is still the primary interpretation, despite being the worst choice. --Tordanik 08:45, 20 May 2012 (BST)
Because the word 'natural' does not fit well with imported types of trees heavily managed for rapid production of products. There is now the tag landcover=trees' ... undocumented .. but I assume this would be used as a general cover anything tag. Warin61 (talk) 08:22, 31 January 2016 (UTC)

The key wood=* is deprecated

The key wood=* is deprecated. Please consider to use the approved key leaf_type=broadleaved/needleleaved/mixed instead of wood=deciduous/coniferous/mixed. Please consider also updating the tagging of existing elements that you have mapped yourself. --Rudolf (talk) 22:15, 6 June 2014 (UTC)

Ancient woodland

What would be the best way to tag ancient woodland, that has never been harvested for anything? Has that been considered before? Pmackay (talk) 16:52, 30 September 2014 (UTC)

natural=wood, additionally, it would probably have boundary=protected_area, and you could add tourism=attraction, or historic=yes if relevant (I'm not sure how useful that last tag is). This, of course, does not not necessarily imply the forest has never been "harvested". There's no specific way to do that, and I'm not sure that's really verifiable (who knows what people 2000 years ago did). -Alan Trick (talk) 17:00, 30 March 2015 (UTC)

Burned Woodland

Fires are a common part of a wood's lifecycle. Shortly after a wildfire, the burned areas usually cannot be classified as woodland, although they will grow back someday. Perhaps we should use burned:natural=forest to indicate that an area has burned, with burn_date=*, burn_type=ground/surface/crown (See Wildfire#Types_of_wildfires on Wikipedia) and whatever tag best describes the current environment. As the environment returns to woodland, keep the burn_date=* and burn_type=* tags to help the user anticipate the age of the oldest trees in the wood. --Amcginnis 20:54, 15 July 2016 (UTC)

Please .. not another variation! Rather than burned:natural=forest use burned:natural=wood ... or burnt:natural=wood. While a burn may have taken place some tress survive fires quite well, so the burn date may not indicate the age of the trees. Many Australian trees are fire tolerant and regenerate even after all the leaves and small branches are burnt off. Warin61 (talk) 00:21, 16 July 2016 (UTC)
If the trees are fire tolerant, is a forest fire significant enough to map? Maybe specify if the trees are fire tolerant if you want to map this, possibly by stating it implicitly with taxon=*. Where the trees are not fire tolerant, burn_type=* can help the user determine if any trees survived the fire. In many North American pine forests, trees rarely survive crowning, but some tall trees may survive a surface burn. Indeed, stating a burn date in revived forest may only be useful when crowning has occurred. To illustrate, here are a few possible uses:
To simply state that this heath was forest that has since burned:
To explain that the heath was forest that burned in July, 2012:
To explain why an area is so thin by stating that some trees were burned:
taxon=pinus ponderosa
To state that the forest was devastated by fire in 1989 and has since grown back:
taxon=pinus ponderosa
In the second example, stating burn_date=2012-07 helps the user predict the condition of the heath. This can be invaluable information if the area becomes neglected by mappers for years; even though the map says the area is a heathland, the user can infer that it better resembles forest. Stating taxon=pinus ponderosa implies that the forest is comprised of vegetation that is not very fire resistant, so information like burn date may be relevant.
Is "burned" American English for "burnt?" If so, then definitely use burnt:natural=wood. Otherwise, it probably doesn't matter as long as everyone agrees to use the same tag.

What is a tree?

When do scrub/shrubs become a trees? I use a minimum height of 3 metres and a single trunk. Warin61 (talk) 05:44, 8 February 2019 (UTC)

Re: >=3 meters That's about what I think as well, though some trees have a trunk that splits just a meter above the ground, so I wouldn't consider that a requirement. But I believe this is regionally variable to some extent. --Jeisenbe (talk) 05:22, 13 August 2019 (UTC)
Wikipedia says "Some definitions state that a shrub is less than 6 m and tree is over 6 m. Others use 10 m as the cut off point." so 3m is probably too short of a cut-off to use. --Jeisenbe (talk) 03:02, 20 August 2019 (UTC)
I would consider <6 m plant that will grow into something that is clearly a tree as a tree Mateusz Konieczny (talk) 09:22, 21 August 2019 (UTC)
That's a good point. Orchard trees are rarely allowed to grow above 6 meters; they are pruned shorter to make it easier to harvest the fruit, but they are still trees. I believe the idea about the 6 or 10 meters cut-off is that a tree should reach that height at maturity in a natural state. --Jeisenbe (talk) 10:04, 21 August 2019 (UTC)

Wood vs Forest

I've removed some misleading sections about the supposed difference between natural=wood and landuse=forest:

"primarily used in less developed or uninhabited regions of the world, often with intact natural environment, such as: Eastern Europe, Latin America or Asia. Almost all wooded areas in these regions are of natural origin; only some wooded areas are managed. "

This is misleading. I live in the most remote part of the Asia-Pacific region, and the apparently wild forests are heavily used by the local people for firewood, timber, hunting, etc. The same is true of the Amazon rainforest: the local people have lightly managed the forest for centuries.

Even apparent wildernesses like Yosemite National Park in the Sierra Nevada mountains of California were managed by humans - the native American people who lived near Yosemite would often set intentional fires to keep the Yosemite valley floor open for hunting.

Furthermore, many of the woodland areas in remote parts of South America and Asia have been mapped from aerial imagery or from imports. When mapping with these methods it is not possible to confirm the level of forestry management for an area of woodland. Please see Forest for more details on the issues.--Jeisenbe (talk) 15:24, 9 June 2019 (UTC)

When mapping from imagery the use of the land cannot be determined. Use natural=wood as this says nothing about it being 'managed' or not nor does natural=* say anything about it being 'natural' or not. Many mappers would not be able to determine if the land is used even from a survey, so the landuse=* should be reserved for those that can determine what the land is actually used for. -- Warin61 (talk 13 August 2019 (UTC)
Note that "landuse=* should be reserved for those that can determiner what the land is actually used for" is not what most of mappers actually do, landuse key is not used only for land use Mateusz Konieczny (talk) 05:46, 13 August 2019 (UTC)
I cannot speak for most mappers, only myself. I would encourage all mappers to only use landuse=* for the land use, and to do so correctly. For example landuse=military should only be for areas used by the military, not for areas of 'danger'. One mapper tagged electrical substations (areas of hi voltage) as landuse=military, it looks pretty... Tagging any area of trees as landuse=forest is not good practice, use natural=wood. If there is an objection to natural=wood then use landcover=trees Warin61 (talk) 06:17, 13 August 2019 (UTC)
There's always debate about how descriptive vs prescriptive this wiki should be. As a small minority of active wiki contributors in a very large, diverse, global project with an "open" data model, I believe we should be humble enough to focus on what the community of mappers are doing, rather than spending too much time telling them what they are doing "wrong". I'll go first, and admit that I find it very difficult to decide when to use "landuse=forest" instead of "natural=wood", in spite of growing up in the middle of a National Forest, where the whole economy depended on forestry of various kinds. Fortunately, openstreetmap's data on woodlands is still very valuable.
If you want my opinion: I would recommend mappers consider adding things that can be observed, like leaf_type=* and leaf_cycle=*, and making sure the boundaries of woodland are clearly marked, by separately mapping natural=scrub, wetland=swamp and landuse=orchard, rather than trying to figure out how to tag management levels of most woodlands. And I recommend against using landcover=trees, a poorly supported tag which solves none of these concerns. --Jeisenbe (talk) 07:00, 13 August 2019 (UTC)
"One mapper tagged electrical substations (areas of hi voltage) as landuse=military, it looks pretty..." - suggesting that blatant tagging for renderer is in any way similar to supposed differences between landuse=forest, landcover=trees and natural=wood is ridiculous (except that main differences that landcover=trees is rarely used and has no support among editors or no/ridiculously tiny support among data consumers). Mateusz Konieczny (talk) 08:28, 13 August 2019 (UTC)