In Wales, large areas of access land and/or registered common land have been tagged with natural=heath. The areas affected are mainly concentrated in the Snowdonia and Brecon Beacons National Parks, and the extent can be seen here: [overpass turbo].
Most of these areas were originally mapped as designation=access_land, reflecting their status under the CRoW Act 2000, which grants a right to roam. In addition, a significant amount of this land is registered common land under the Commons Act 2006. These areas should also carry the tag designation=common. The wiki notes that designation=common should be paired with either a leisure=* or landuse=* tag to describe the primary use of the common.
I don’t think there’s much doubt that applying natural=heath to such vast areas is incorrect. The challenge, though, is how best to correct this. Simply removing the natural=heath tag would leave very large areas blank on the map. Since many of these areas are grazing commons, my current thinking is to use: designation=common landuse=meadow though I’m not entirely sure this is the best choice. It has been argued that landuse=meadow is not a good fit for the land covering however I would argue that the tag should only mean “animals graze here” and nothing else.
I’ve already tried this approach on two smaller commons: Gelligear & Merthyr Common And Rhymney and Bedwellty Common. In both cases I tagged the area with landuse=meadow. For Rhymney I left it at that, while for Gelligear I also mapped smaller patches of natural=heath within the common.
The Brecon Beacons Common looks like another strong candidate for this approach, but I’d appreciate feedback before making broader changes.
Looking at an area I’m familiar with (Gower), I do agree that not every common is 100% heath, as there are some marshes, grassland and scrub in places too - so i wouldn’t say they’re 100% meadow either. I also think that meadow should just mean “animals graze here”.
I would argue that, as soon as CRoW / Common land are complex enough to need defining by multipolygon relations, the relation shouldn’t contain any physical descriptors at all (e.g: landuse= or natural=). There’s no reason I’m aware of for the legal access designation being linked to the physical characteristics.
For at least a few of the examples of natural=heath in Wales it appears that this was used initially to show physical characteristics and then adopted (I believe incorrectly, and lazily!) by later mappers to add access= and designation=* details.
As for using landuse=meadow as you have started to, I’m afraid I disagree with this here. I think that natural=heath is a pretty spot-on description for this landscape, varying in places with perhaps natural=scrub;fell;wood and landuse=forest etc. A meadow, to me (and it seems to the wiki), is a cultivated (hence landuse=) area of land with mixed plants for feeding animals (either in-situ or after harvesting)
On the plus side, thanks for bringing this topic up again - it’s brought my attention to just how much of Welsh countryside isn’t mapped at all. I feel a little project coming on…!
while leisure=common does appears to be the correct tag for this situation originally it now appears very controversial and seems like its going to go the same way as leisure=nature_reserve where is being slowly phased out for boundary=protected_area
as leisure=common wiki states a good point too that a common is a legal concept more than a physical so with that in mind describing the common with designation=common would be sufficient however although a common itself is a legal concept there is still physically a fence area of land that has a name that needs a generic tag. This land also has animals roaming freely which i think should be tagged in some way for users such as motorists.
@mstrbrid I can see your point of view but as I just explained I think it does need something generic to describe the named land rather than the common. The main reason I suggest landuse=meadow is because the lack of an alternative at the moment, I have suggested possible alternatives in the past like landuse=common or grazing=true, something along those lines to indicate that this is a area where animals roam freely for users like drivers.
I’m not sure if your referring to a specific example here but in general I don’t completely agree. In my experience while there are large areas of heath, there are also large areas of grassland, much of this hasn’t been mapped due to how difficult it is to distinguish grass from heath on imagery.
I was thinking about one of your examples Rhymney and Bedwellty Common, but it would apply to most of the Welsh natural=heath that I looked at. As for distinguishing between different landcovers, why is this situation different from mapping anything else remotely? If you don’t know what it is, only map it to a level of detail you can be reasonably confident of from the sources you have access to. For mapping large areas of landcover characteristics, perhaps Sentinel-2 imagery would be useful? (If you don’t already use it)
Would you mind clarifying what you mean to achieve here with the “named land”, and what “named land” is? If: -
it’s that an area within the designation=common has a distinct name, then a place=* node for undefined areas, or a type=boundary relation for clearly defined areas seem appropriate, or;
it’s that you want to describe the ground cover of a defined (or named) area, then I’d suggest…don’t do it on the same object - at least not for areas you’d like more granular landuse=* and natural=* detail. What you’ve done with the Tremains Wood LNR recently is precisely what my first post was advocating for; designation & access on a boundary and physical geography on the closed ways / multipolygons.
As for highlighting that grazing occurs in the area, this would be compatible and appropriate on the boundary object where you’ve got access tags. To highlight the risk to motorists, then I would propose hazard=animal_crossing along the appropriate highways as being suitable where it’s signed at the entrance to an area.
I’ll let others come in with their tuppence-worth for a moment now…
Why? Other than tagging for the renderer, which we don’t do. Why does it need an additional tag if there’s a perfectly good leisure tag available for use?
If you want to use landuse=common as well that’s seems perfectly fine too, there are currently 271 examples of that knocking about in the database.
There’s always place=locality if there really is no more descriptive tag for a named location.
I suspect that the challenge here is that there won’t actually be a correlation between any of these, requiring separate features to carry each tag, As an example, this is the Natural England / Defra map of (and I’m quoting here) “CRoW Act 2000 - Access Layer - Shows land mapped as access land under the Countryside and Rights of Way Act 2000 (CRoW Act).”. Looking at that northeast section, the common area is much larger that the NE / Defra map shows. The whole thing is common - so says the law available here and on boards around the area.
The bits that are heath and the bits that are grassland are mostly tagged as such, but the natural=heath on the whole area is a hangover from when it was first mapped (without detail) many years ago.
I think that the approach here, and in other bits of Yorkshire and also in Wales will be similar - add the detailed landuse/natural tagging and then leave the “large polygons” for designation=access_land or common as appropriate.
One other thing - in many cases the old “broad brush polygons” can be aligned to more recently added walls and fences. If adding detail in upland areas I often find myself doing that too.
Sorry for my lack of reply, life caught up with me but it has given my time to think and consider your comments.
Up till now i’ve been using landuse=meadow to describe any fenced grazing land. After taking your comments onboard I now feel this was a mistake and should be used for maintained grass grazing fields instead. likewise If a gazing field is made up of heathland it should be tagged with natural=heath. With this conclusion I accept that landuse=meadow is not correct for commons.
Sorry, i’m been struggling to find the right way to word this to get my thoughts out and i think i failed so far. Let me start again focusing specifically on Rhymney Common. There is no doubt that it should be tagged designation=common name=*. however my issue with this is that i think designation=common is too specific of a tag to be used as the main tag.
leisure=common can be added but there are issues and it doesn’t really solve the problem. leisure=common is basically just a older, misused version of designation=common. it is practically just as rare too. designation=common has 600 uses and leisure=common has 800.
The last option i can think of which @InsertUser also suggested is place=locality which I do think fits but it does feel like a its just a fixme waiting for a better main tag to come along.
This is a little border line mapping for the renderer but heres my logic. the name of the common should be rendered however looking at the data from a renderer data consumer’s perspective. designation=common is far to niche to consider. if you were to then you would have to consider 1000s of equally niche tags all over the world. so i’m just looking for a more generic way to say “this area of the world has a name”
On a side note it would be nice if place=locality areas rendered in carto but that’s a separate conversation.
edit: type=boundary boundary=place place=locality relation is also a option as @mstrbrid mentioned.