Talk:Tag:man made=embankment

Cutting

What about the opposite where a road etc is dug down, or rock blasted away to make a sort of narrow man-made valley ? MortenLange (talk) 21:51, 7 November 2016 (UTC)

See Key:cutting --Klumbumbus (talk) 21:57, 7 November 2016 (UTC)

Yup, thanks. I saw that some minutes later and thought I indicated it here, but somehow that edit failed :-) MortenLange (talk) 14:23, 8 November 2016 (UTC)

Area

Shouldn't man_made=embankment be valid also for areas ()? --NonnEmilia (talk) 09:10, 25 November 2016 (UTC)

The short answer is NO --geozeisig (talk) 06:11, 26 November 2016 (UTC)

LoL. Where I can read about a long answer?

The question is: how can we tag a wide and complex embankment/levee (with many levels and a lot of different features on it like roads, paths, cycleways, benchs, ...)?

I'm talking about the complex levee system (riverbanks for flood control) along the Po river[1]. --NonnEmilia (talk) 07:59, 26 November 2016 (UTC)

Use man_made=embankment on a way where the cliff is. It is not so good to use a closeway because it will not be rendert. My be an examble is here --geozeisig (talk) 10:39, 26 November 2016 (UTC)

Now it's rendered. See https://www.openstreetmap.org/way/507360111 for example. I still fail to see why it can't be used on areas. --naoliv (talk) 18:46, 20 December 2017 (UTC)

Maybe you have to check the right direction.--geozeisig (talk) 07:34, 21 December 2017 (UTC)

"Maybe you have to check the right directio" - I do not understand at all what you posted Mateusz Konieczny (talk) 09:45, 21 December 2017 (UTC)

We have clockwise ways https://www.openstreetmap.org/way/507360115 and anti-clockwise ways https://www.openstreetmap.org/way/507360111 They are both closed and rendered. We also have +6000 closed ways with embankment=yes or man_mane=embankment and I still fail to see any valid reason to disallow the usage on closed ways. --naoliv (talk) 10:54, 21 December 2017 (UTC)

In my interpretation https://www.openstreetmap.org/way/507360111 is a line, just closed one Mateusz Konieczny (talk) 09:45, 21 December 2017 (UTC)

Makes sense, indeed --naoliv (talk) 11:06, 21 December 2017 (UTC)

You may call it closed way but I think it can be also interpreted as an area and this is exactly what OP wanted to do as an area. So maybe there is some misunderstanding. Also this example demonstrates that the closed line obviously is rendered.. and if it isn't in some renderer it should be. RicoZ (talk) 20:31, 25 October 2019 (UTC)

It's rendered as a linear feature. Almost all lines, like railways and roads, can be drawn as closed ways when they make a circle, but that does not mean they should be interpreted as areas. --Jeisenbe (talk) 23:31, 25 October 2019 (UTC)

Correct. In this case there's the area=yes tag. It was created to distinguish a closed way from an area because both are drawn the same.
So a closed line with man_made=embankment and area=yes could be acceptable? --NonnEmilia (talk) 07:38, 28 October 2019 (UTC)

You could do this, but it is less helpful than you would think for database users, since closed ways can still be ambiguous when a tag is allowed to represent both an area and a linear feature. I don't see the benefit from considering an embankment an area. It's well-defined as a linear feature with the high ground on the left-hand side and the low ground on the right-hand side. If you really wanted to map the whole area of a levee or dike, they you would probably want to include the slopes on each side, right? In that case, it doesn't work to use man_made=embankment, because this tag is already defined as describing the high side of the embankment, not the low side. Perhaps you could try a new tag like "area:man_made=embankment" or "area:man_made=dyke" (depending on the feature) which can be used for this purpose? --Jeisenbe (talk) 11:54, 28 October 2019 (UTC)

Still wondering which area OP wanted to map:

• the area delimited by an embankment on all sides - this is already possible with a circular way and doesn't need area tag
• the area of the sloped terrain, the embankment itself - because of the directionality of man_made=embankment afaics you can't do that, there is an upper and a lower end of the embankment which can't be represented like this. RicoZ (talk) 17:14, 1 November 2019 (UTC)

What's the status for areas in 2023?

• Main article says "This tag may be used on open ways (open polylines) way as well as closed ways closed way (closed polylines)."
• But "infocard" on the right restricts areas, and also JOSM says it's not applicable.--UrSuS (talk) 06:54, 16 March 2023 (UTC)

Earth walls / barriers

Previous discussion: https://wiki.openstreetmap.org/wiki/Talk:Key:barrier#soil_barrier_.2F_Earth_Wall.3F

I want to ensure that this specific feature is better documented in the wiki since there doesn't appears to be clear consensus how these "barriers" should be mapped. They are 100% barriers to keep people out / animals in and are usually formed using soil from the immediate area. In Ireland (and other parts of the world) they can be found faced with local stone too.

Current guidance (and support in editors) pushes map editors in the direction of man_made=embankment and embankment=yes but the relevant wiki pages are almost solely focused on embankments used in transportation (roads, railways).

Questions:

• Is the guidance above correct?
• Are additional tags appropriate (height=*, etc)?
• What changes are needed to the relevant wiki pages to improve them?

--Dónal (talk) 12:17, 27 November 2020 (UTC)

As sound barrier

Can I use this to map a sound barrier between a large road and residental area? I'm a bit unsure whether an "embankment" need to have a road or railroad on top of it, or if it just can be a low and very wide man-made hill. --Christoffre (talk) 10:22, 22 February 2022 (UTC)

Definition right now seems to very clearly include such use, and I can confirm that it is also used in this way @Christoffre: Mateusz Konieczny (talk) 13:25, 22 February 2022 (UTC)

Thanks Mateusz Konieczny Christoffre (talk) 11:44, 25 February 2022 (UTC)

Actual usage

The section Actual usage, ambiguity and controversy" has nothing on actual usage! I think from the discussions emerges that actual usage is for the main part to map the top line of a man-made embankment slope, which is compatible with the current wiki text. I think it is hard to give numbers or percentages for different usage, e.g. based on older versions of this wiki text, or because of lack of a better altenative. Still, maybe some estimated statistics are available somewhere.

Man-made or Natural?

I'm currently mapping a small area that used to a garbage dump 40–60 years ago. Today it is mainly a forest, but also a ski slope and recreational area.

The steep earth banks here look all natural. But with my knowledge about the local history they are obvious old man-made garbage piles. Thus, I have mapped them as man_made=embankment. But according to the policy "map what is on the ground" I should actually have mapped them as natural=earth_bank, as that is what they look like.

What is proper? Embankment or Earth Bank? –Christoffre (talk) 22:37, 25 March 2023 (UTC)

A proper key for this area may be Tag:man_made=spoil_heap or abandoned:landuse=landfill, I think.--ManakaAo (talk) 02:18, 23 April 2023 (UTC)

I'd agree with ManakaAo about using a lifecycle prefix for mapping the area as abandoned:landuse=landfill to convey that aspect. From your description I'd add natural=earth_bank to the edges, just like I add natural=cliff to historic quarries around me. --TrekClimbing (talk) 06:21, 23 April 2023 (UTC)

Area of the embankment slope

I think usage of man_made=embankment as the crest line of one or both sides of an embankment is well established. The directionality (right is low) makes it hard to use it for an embankment toe line or full outline of the embankment (because the toe line by definition has no lower side, and because if you draw one slope as an area, the direction of the toe line will show the upgoing hill as down-going, effectively mapping a depression in the ground instead of one slope. However, there is a way to do it: use the embankment crest way as an outer in a multipolygon, then draw an untagged way to complete the area of the slope, and make that an outer as well. The MP is a valid area, and can be tagged with natural=grass, or other applicable tags. You could also tag it as e.g. man_made=embankment_slope. Data users could improve the current tagging using the extra geometry information, where they can see what the upper side of the slope is and what the extent is, and any extra tags for the area (it might even have a name). E.g. They could decide a nice colour or shadow striping style to show the high side and the extent. Data users who do not incorporate this in their style, will show exactly the same rendering as before.

I have tried this, and it works fine with current renderings, because there is nothing new. --Peter Elderson (talk) 19:36, 8 September 2025 (UTC)

The drawbacks I see using the crest line as part of multipolygon area to indicate the slope:

* The crest is in practice almost never a line but also an area, it makes sense to map the crest as area if you are mapping the slope as area and then this trick does not work.

Firstly, I don't think it's a "trick"; it's regular use of mapping an area as a multipolygon. In this case, the crest line actually is part of the area outline, so it's correct to use it as an outer. Secondly, in my experience, in OSM man_made=embankment is used as a line element to represent one (section of) slope. I have encountered (and mapped) crests with man_made=embankment all around, but it's an exception to find a stand-alone dyke with slopes at all sides. Never mind; if you want to map the slopes of such a dyke, you just cut the man_made=embankment way and map two slopes which happen to meet each other at the ends. If you want to additionally map the full crest as an area, you can use the two crest ways as outers in an MP, or just retrace the outline to add a polygon, just like when you add a meadow between canals and ditches.--Peter Elderson (talk) 07:41, 9 September 2025 (UTC)

* Using a multipolygon makes the things complex

True, MP areas are more complex than simple area polygons, but they are used in abundance. In this case, they are not monster landuse MP's or monster waterway MP's, just single areas with the upper side as one outer and the rest as a second outer to complete the outline. --Peter Elderson (talk) 07:41, 9 September 2025 (UTC)

I still think things are too complex. Suppose you want to map the area of the crest and both slopes. That would mean 3 MP's and 6 ways, for my proposal just 3 ways/area's -- Emvee (talk) 20:41, 10 September 2025 (UTC)

Mapping one crest line with a way is pretty much established as the starter's solution. One step further, crest and slopes can be added as simple natural or landuse polygons, but the slopes would not have an upper or lower side. Mapping the slope area as an MP with two outers is not a big step. I think it is even simpler than an MP with an outer and an inner! For the crest area, if applicable, this is not needed, because it does not have an upper or lower side. I know many current mapping practices which are much more complex than this; an MP of just two outers is about as simple as you can get (though I sometimes map MP's consisting of only one outer). But remember, I am not advocating all embankments should be mapped like this; I am just saying there is a way to do it with regular methods; nothing new, breaks no existing mapping or data use.

--Peter Elderson (talk) 21:47, 10 September 2025 (UTC)

* When using an area, a way to indicate with side is up could be to agree that the line formed by first(/last) node and the second node is the side that is up. Area's in OSM are build up of nodes, for example n17,n52,n82,n22,n17. For this area, the line formed by n17,n52 would the high side. What I like about this is it is complete self-contained, the drawback is that some additional editor support would be needed. Emvee (talk) 21:04, 8 September 2025 (UTC)

(Hm... How would the data user know at which node the crest stops and the side of the slope begins? Direction change of the crest line is to be expected, so you can't use that as an end-of-crest-marker.) I think that this would be an innovation which could work, but needs development of tooling and data using applications. I think the MP solution uses existing mapping, tagging and tooling, in itself changes nothing for data users, it just offers an opportunity for refinement of rendering. --Peter Elderson (talk) 07:41, 9 September 2025 (UTC)

I think my explanation n17,n52,n82,n22,n17 above should be clear. Thinking about it it is only working for a simple case where the slope is mapped as area with 4 points and that is not generic. I think you could make it still work by specifying that the first nodes touching another area (the crest) are "up" but that is not such a nice solution in the sense that not all information is self-contained. Instead we could use the three ways/area's with one MP that has these three ways/area's with roles crest, landside_slope and waterside_slope. -- Emvee (talk) 20:41, 10 September 2025 (UTC)

That could work, but it is a new type of relation, and new tags, and it requires complex handling by data user, compared to regular ways and areas with their own tags. The high end of the slopes is not defined by the role in the relation, so, for e.g. shaded rendering, it still has to be calculated from the surrounding landuse/landcover, which may or may not be present. I don't see this happening any time soon! In discussions, various ingenious possible solutions have been mentioned, but none have gained traction. That is why I presented this limited solution to map a (section of) embankment slope using regular mapping: crestline as a way, which is an outer in a regular MP area. If a mapper encounters a mapped crest line, (s)he can add the slope to it simply by adding the missing outline of the slope, and put the two ways as outers in the area MP, then add any the tags for the area. No special rendering required, but renderers have the option to do something with the slope. For innovation, an optional incline=up|down|degree tag could be set? That would require incline to be extended to areas with an embankment outer, e.g. incline=down means down from the embankment outer. --Peter Elderson (talk) 07:08, 11 September 2025 (UTC)