Desert in the south of Israel

You may know that I’m a fan of the landuse definition as I like the map to be alive and show farmland, forest and other elements to display that we are a living and active country.

in Zoom level 8 you see forest
in Zoom level 9 you can see farmland
in Zoom level 10 you can see nature=sand

Because a huge part of Israel is desert it would be interesting to show and represent that in OSM.
So I added the southern Negev and Arava Desert (http://www.openstreetmap.org/browse/way/142100443)

There is actually no nature=desert approved (http://wiki.openstreetmap.org/wiki/Proposed_features/Deserts) so the question is, how to tag a desert?

Question:
I did it by nature=sand (to show it in yellow) but this could maybe a conflict as not every part of the desert is sand.
Maybe it could also be a problem to have everything in yellow in the south of Israel.

What do you think?

Most of the Negev is not sand - it would be very misleading to mark it as such.
Hammada is the most common formation.
There are many official tags for arctic or temperate climates (Heath, wetland, tundra etc.) but none for deserts, and this should be corrected.

I can tell you the truth. I would say that I’m a very technical person.
With programming skills and a high IT knowledge but I don’t have a clue how OSM is Proposed_features. In terms of usability it is a disaster.

It seems to me that people would have to open the “http://wiki.openstreetmap.org/wiki/Proposed_features/Deserts” page and than actually edit the wiki page and add their name and if they want to have this feature or not.

Is it really like that? How many people do we need to get this accepted?
Should we maybe start another try to get this accepted? How many users do we need?

BTW: I have also used the OSM help system to ask the same question.
Currently only the suggestion to just use the desert tag.
http://help.openstreetmap.org/questions/9586/how-should-i-tag-the-desert-in-south-of-israel-jordan-egypt

@Mr_Israel, I understand your desire to eliminate “white areas” - this is what I want too.

The old/new question is, do we tag for the renderer? I agree with valleyofdawn that tagging a vast area such as the Negev with natural=sand is simply incorrect (actually there are lots of sand in the beaches along the Mediterranean coast). There is apparently a proposed desert=hammada tag.

I don’t think we need to wait for a feature to be approved in order to use it.

On the higher level, I would like to see in OSM all the main geographical regions/areas: Galilee, Golan, Carmel, Coastal Plain, the Central Mountain Range (Judaean and Samarian hills), Jordan Valley, Judaean Desert, the Negev, the Arava. On the second level we would have Lower/Upper Galilee, Yizreel Valley, Sharon…

I believe there are solutions for all the above in terms of tagging (see here) , but again I don’t mind too much about that as long as there is no alternative.

See also landcover tag.

Ok.

I agree with your opinion that natural=sand is wrong and changed it to natural=desert.
The result you can see here => http://www.openstreetmap.org/?lat=30.75757&lon=35.32343&zoom=15&layers=M

The area is not getting yellow painted as for sand, but a kind of brown for the desert tag in Mapnik.

I personally don’t like that kind of color as it is way to dark, but it looks as the right definition.
Let’s see till what zoom level this will get rendered.

BTW: If somebody could be so kind and add the Hebrew namings for Southern Negev Desert and Arava Desert to the ways in the south.

Regarding the geographical regions/areas just try to add it and let’s see how this get rendered.
For sure we could do something like this on our own OSM server.
Those kind of things are not flexible and not change the next years.

We need to change two things for the desert rendering in Mapnik:

1. The color (currently brown)

=> Maybe we could change that for our osm.org.il server and later see how it gets rendered the best way and then send as change request to the mapnik team (or by ourself).
I would pick something yellow and a bit lighter than farmland or sand.
Maybe something like this yellow color: www.colorcombos.com/FFFF99-hex-color

2. Adding a name tag to the desert in a low zoom level.

=> The OSM.xml defines actually how the desert should be rendered in mapnik.
http://trac.openstreetmap.org/changeset?old_path=%2Fapplications%2Frendering%2Fmapnik%2Fosm.xml&old=24922&new_path=%2Fapplications%2Frendering%2Fmapnik%2Fosm.xml&new=24922

A example with a text information and the new color
(I’m not aware how TextSymbolizer gets rendered and what all those options are)
###########################

[natural] = 'desert' &maxscale_zoom8; #FFFF99

###########################

What do you think?

I have some further objections to this desert tag.

  • The Negev is now by far the most noticeable feature in Israel, even though it is of little interest to most map users. All the tag tells you is that southern Israel is arid (a very well-known fact). You can’t even see the name “Negev” at any of the resolutions I tried.

  • The desert polygon has a well-defined edge, while in real life deserts do not. Zooming in to (for example) the city Rahat, one sees that the brown area ends just to the south of it. It makes you think that there is a noticeable natural change just south of Rahat, but in real life of course there isn’t. In contrast, other examples of land use which ARE tagged on OSM (forest, farm, urban area) have well defined boundaries (trees, uniform area growing one crop, buildings) while deserts do not.

  • Sinai and southern Jordan are arid like the Negev, but are not marked. Marking them too would not solve the problem, since the desert really continues through all of North Africa and the Arabian peninsula.

  • It is incompatible with the rest of OSM - virtually no deserts are mapped worldwide.

I suggest the Negev (and later the Galilee, Shfela etc. for anyone who wants to add them) exist as a “place=region”, or some similar tag where the data is present for searching, but does not have visible edges or coloring. The way it is now, I think is distracting and misleading, and changing the color is not enough to fix that.

Thanks

@eric22

I fully agree with your points. I’m with you. Thats why I think we should remove the desert in the current way it is displayed.
BUT, before we do that we should ask mapnik to reconsider the way a desert gets rendered today.

And this is what I did now. I opened a ticket asking to change the way a desert is displayed today.
https://trac.openstreetmap.org/ticket/4430

You are welcome to add your feedback and thoughts to the ticket I opened.
BTW: the userlogin is the same as in Openstreetmap. So no need to create a new user.

If I will have time I will create a new proposal for the desert feature.
It would be great to get some more input for this feature request.

Thanks.

Seems like they have removed the rendering completely in Mapnik.

The tagging of vast areas in the south is back to “Sand”.
These are crude polygons with no relation to the situation on the ground.
Most previous commentators agree that this was a bad idea.
I suggest removing all of these areas, except for חולות חלוצה
If there is no objection I’ll do so in the coming days.

+1

I’m ok if you like to remove it.
If most people are not liking it, I have to accept it :confused:

This open resource lets you assess which areas are sandy
http://soilgrids1km.isric.org/