The title is what’s being observed, the most notable old mapped basins that have natural=water + landuse=basin. This one I know get’s flagged by I think Osmose and possibly ID Editor, but only when one mods the object proposing to upgrade the tags.
Created a JOSM toolbar button to search the larger data area (including edits in progress) and found half a dozen and cleaned them up.
If it’s an absolute I’d propose this to be made a (general) validation check. (May exist but not encountered AFAICR).
I’ve always looked at tree covered area with an economic viewpoint and this got written in the OSM world
" landuse = forest is used for maintained or managed woodland. This approach views most woodland as managed or maintained especially in areas such as Europe. natural = wood is used for ancient or virgin woodland, with no forestry use. Oct 5, 2022"
Is there any legit reason why there can be both landuse + natural on the same area?
Just do an overpass search and have a look at some of the values. If you look at the combinations some I’m sure won’t make much sense (but will probably need a survey, rather than guesswork based on aerial imagery, to fix). Some, however may make sense. Near me, this one somewhat does.
I’ve joked before that Osmose “is usually wrong”, but to be fair to the developers of that the front page does say “In no case Osmose-QA should provide you the absolute right way to map, always keep a critical eye”. Unfortunately some mappers ignore that and randomly click to “fix” stuff.
There have been issues there, too. Again, mappers need to look at the actual changes suggested, and at the object being modified, and not blindly hit “fix” on lots of objects in an area. It will break things.
“Harmless” or … I don’t know, can’t judge, lets tag both. (Seem to remember a RFC proposal related to forest that got withdrawn put back on design board’. Might have effect on this and maybe zillions of plots in OSM)
A residential zone under construction what used to be a scrub area. Can only guess the natural tag was left for when construction has completed.
Any other examples please? A validation rule would at least actively call for a review, but only when you wiggled nodes of the object. We can always decide to dismiss.
(PS The big time waiver clause of Osmose does get long in the tooth, for me at least, it’s presented with qualification in JOSM too on each upload prep when “error, warning, maybe” cycle gets called)
No! It appears I have to add some more characters, so …
As Mateusz says landuse=forest, natural=wood is one example (and the wiki description about natural=wood == Urwaelder is totally artificial, the difference is useless because you can’t actually tell unless you do an on-the-ground survey of the vegetation, and also the definition is different in Britain & Ireland, being woodland which has existed since before 1600), but a brownfield site which is getting scrubbed over would be another.
I recently noticed that where landuse=reservoir has been changed to natural=water + water=reservoir that it is now not possible to unequivocally identify reservoirs because tagging is often incomplete in the second case. It’s also not so easy to identify natural lakes which have been enlarged by damming an outlet (Thirlmere in the Lake District).
@SekeRob : the area around Ova Spin is on the edge of the Swiss National Park (there may be a buffer zone now). The area up to the Pass dal Fuorn/Ofenpass was largely deforested in the 19th century for lime making (good documentation in Nationalparkhaus Zernez). However, all the woods one sees now have regenerated naturally from inaccessible remnants, with much of the spread being by Nutcrackers/Tannenhäher, as shown, at this location, in David Attenborough’s Private Life of Plants. Of course woods in the National Park are managed in the sense that rangers patrol regularly to ensure people aren’t interfering with nature off the marked paths. All the woodland in the park should be regarded as natural, and I suspect a fair bit outside the park too. From what I remember the expectation is that it will take 150-200 years before these woods reach their full natural potential, mainly because there needs to be a wide range of age classes of different trees.
There are some reasonable historical photos from the ETH Archives on Wikimedia Commons.
According to Taginfo there are 365k landuse + natural objects, which account for 0.87% of landuse=* and 0.57% of natural=*. We can do further research to see to which values it applies most often, but these low percentages indicate that the combination of landuse with natural is non-standard.
Also to add another example (which I personally don’t really like, but haven’t found a better alternative tagging) that might seem counter intuitive (and, if one follows the wiki to the letter, actually outright wrong):
I sometimes map landuse=meadow+meadow=pasture+natural=wood (or, a bit more commonly, a separate natural=wood area overlapping a landuse=meadow+meadow=pasture) for cases like this (to the right of the road):
There’s sufficient trees that one could say to be standing in the woods/a forest, but the landuse is definitely a pasture and there’s no good alternative tagging (that I’ve found, only an abandoned proposal) for a wooded pasture.
I don’t see any evidence to grazing. The field layer is quite high and there’s a suggestion of the presence of shrubs & saplings. I do agree that it is a kind of transitional habitat though.
You can see the fence in the photo, but apart from that I also know from local knowledge that there are often horses grazing there. But they move the animals between different pastures to let the grass regrow.
I am late to the party and thread is tagged as answered. But from my perspective landuse and natural should not overlap itself and vice versa.
Originally these two hierarchies were about used, or not used larger areas. So if some place is beeing actively used its landuse, if its left by itself its natural.
This has, over the past 15 years, lost its purity. People have added micromapping landuses like flowerbeds and abused both hierarchies for micromapping and layering. This is IMHO mostly caused by landcover not beeing rendered, which would be the best solution for micromapping within a landuse or natural. But - if its not rendered its considered bad.
For me landuse/natural overlaps are a semantic error which i fix whenever i stumble on it.
I have create a debugging view for that in my QA tools for most of Germany - This is Northrhine-Westfalia:
