The wadies I have encountered in the desert regions of the western US states seem all to have ephemeral borders. Every time water flows down the "border" changes. Also the center line in most cases is difficult to determine. BTW many of thr rivers here in Northern Italy that have their springs in the mountains and their very wide riverbeds in the nearly plains in the wider Po valley. The material here is pebble, and not sand ad in the desert-set wadies, but zhe variability of the shape of the riverbed, and of the water course are similar. They are not called wadi, and I don't know the proper term.
On Sun, 26 May 2024, 17:11 Kai Johnson, <b.tw...@gmail.com> wrote: > > On 25 May 2024, at 22:38:20, Peter Elderson <pelder...@gmail.com> wrote: > > > > I like this proposal very much. > > Thank you for the review and the comments. > > > One thing: by its nature, other naturals will overlap with natural=wadi, > right? Any thoughts on that? > > I think that's a fairly common thing with features tagged with natural=* > and > we have some reasonable ways to handle it. > > First, where natural features overlap, their boundaries don't always > coincide. In this case, the features can and should be mapped as separate > elements. Mapping natural=scrub in an arid environment is a good example. > There may be scrub in a wadi, but if so, there is likely also scrub in the > areas outside the wadi's banks. So, the natural=scrub area need not share > the same geometry as the natural=wadi area. > > Second, if the boundaries of the natural features do coincide precisely, > the > natural tag can have multiple values or the features can be mapped as > separate ways with shared nodes. I don't know that this would be common > when > mapping wadis, but as a separate example, we use coincident ways when > mapping sea stacks where the natural=coastline and natural=bare_rock > polygons coincide but the two tags cannot be placed on the same element. > > Third, one of the defining characteristics of a wadi is a bed of > unconsolidated and often poorly sorted alluvium. That is, sand, gravel, > cobble, boulders, or the like. This can be mapped using surface=* in > combination with the natural=wadi tag. > > > And, you say you can map it as a node. That seems a bit strange, because > it's such an elongated feature. > > I don't think that's the preferred way to map a wadi, but just as > natural=valley can be mapped as a node (per the wiki and established > usage), > it could be acceptable for natural=wadi. As with a valley, even though the > banks of the wadi may be distinct, the start and end of the wadi may not > be. > Rather than making an arbitrary decision about the boundaries, a mapper may > prefer to place a node at a central location to map the wadi. > > To add to that, many wadis have well-established names and are important > landmarks. So, placing the name on the map with a node is valuable even if > the boundaries of the wadi have not been mapped. > > -Kai > > > _______________________________________________ > Tagging mailing list > Tagging@openstreetmap.org > https://lists.openstreetmap.org/listinfo/tagging >
_______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
