I’m currently mapping the Nevada desert but its a bit tricky. Do you map the soil or what is on the soil? Most of the soil is either desert or scree, but on top of that there is always scrub. Ultimately I would love to use desert+scrub, it represents landscape as it, but it doesn’t render on OSM map.
So any suggestions? At the moment I’m doing soil, so lots of scree on hills but they look bare which they are not.
Hi! Good to meet a fellow Desert Southwest mapper!
First off, be careful with natural=scree. That’s specifically for broken, angular rockfall from adjacent rock faces. Most of the loose rock in our deserts is unconsolidated alluvium. That is, rock that has been eroded and transported by water (or wind sometimes). Definitely use natural=scree where it fits, but consider natural=sand/gravel/cobble/boulders (from Grain size - Wikipedia) for bare loose rock or soil from other sources.
All that is for bare surfaces without plant life, though. If the area has plants in our deserts, it’s probably natural=scrub. That covers growth up to small trees, so mesquite, ironwood, acacia, etc. still fit.
Some areas of the Nevada desert could qualify as natural=grassland even though the grasses are rarely green. But if the surface doesn’t have mid-sized scrub or small trees, it could be natural=heath.
Be careful with natural=sand too. There are a lot of old polygons around with this tag and it drives me nuts to go out there and see that the entire area is covered in creosote bush. That’s natural=scrub not natural=sand. It has to be bare sand, like sand dunes, for it to be natural=sand. And if there are actual dunes, you can use landform=dune_system as well.
I’d be careful with natural=desert as well. That tag has some issues. Like, what is a “desert” and where are its boundaries? And, could the area be better mapped using more specific tags?
If you’d like more detail in your desert mapping, consider natural=wadi(proposed) for broad washes or natural=desert_pavement (experimental). And you might look at the Glossary of Landforms for more ideas.
I come to the conclusion that I did a lot wrong, but luckily you can change the type of polygon with 1-2 clicks
I will check my updates. I think ultimately OSM falls short in mapping dry areas. We need dry scrub and rocky scrub, because the current scrub has no link to anything dry. You might think the desert is full with life and vegetation, which is not.
Mapping things one way and coming back to change them when you find out more is good! That’s how the map gets better!
And I totally agree that OSM falls short in mapping dry areas. Many of the tags that make sense aren’t documented or even widely used, and few of them are rendered. On the other hand, we have atrocities like Sarcobatus Flat that are tagged and rendered as sand, but clearly are not (as shown below):
In my opinion, none of the OSM defined land cover tags come close to describing the arid and semi-arid landscape in the deserts of the Southwestern US. And probably for similar climatic areas elsewhere in the world.
Regarding natural=desert_pavement, at least in the area I grew up there were lots of scattered creosote bushes, foothill palo verde trees, cacti, etc. throughout the area where the top surface of the soil was considered to be desert pavement. The amount of vegetation varied dramatically between areas of desert pavement. And there could be rock outcroppings, etc. as well. So a single natural=desert_pavement tag would not be sufficient.
I would lean toward some scheme where the density of vegetation and its average height could be specified.
If someone is smart enough to come up with decent tags and persistent enough to get them accepted by the OSM community it would be wonderful. Unfortunately, I don’t have those skills.
For an example, here’s my attempt at tagging landcover in Joshua Tree: Relation: 15533394 | OpenStreetMap. I went natural=scrub + scrub:density=sparse and a thorough description.