From OpenStreetMap Wiki
Jump to navigation Jump to search

discuss Tag:natural=bay:

How to map? point of this tag?

I'm confused about the point of this tag

Can you write a "How to map" section. What's the idea? draw a way around an area of water? Would this way also be tagged natural=coastline or natural=water? Or is the idea similar to Key:place, to have a node naming thing bay?

-- Harry Wood 20:02, 30 May 2010 (UTC)

Well, AFAIS it apply to both node and area. Regarding node, usage is pretty obvious; regarding area I'm not sure about this, maybe we should eventually discuss it at mailing list. Yarl 15:26, 1 June 2010 (UTC)
I'd like to know about this as well. Near me there are several "bays" and "coves" which are already drawn as water because they are enclosed by natural=coastline ways. Right now there are nodes in the middle of each bay tagged natural=bay, however since several are quite large I'd rather map them as areas. Should I just create an overlapping area (or multipolygon)? It doesn't quite seem right, but neither does pushing the coastline out. -- Joshdoe 13:42, 28 July 2011 (BST)



Is adequate the use of bay for cove [1]. In Spain, there is a clear distinction between cove and bay. The first has dozens or meters of witdh, but the second is very large. I don't feel confortable to tag more calas (coves in spanish) as bays. I think perhaps we have to create some other tag for include both. What about natural=water_entry and water_entry={bay, cove, ...}?? Do you know it's useful? Thanks in advance,--Xan 17:55, 8 August 2010 (BST)

At the larger end of things consider also a "bight".
-- Hamish 10:05, 26 August 2012 (BST)
Presently, the coves in my area have been tagged as natural=bay
-- mtc 12:00, 28 April 2017 (EST)

What about straits?

How would we tag a body of water that connects two larger bodies of water, or a body of water that separates an island from another body of land? -- DENelson83 00:19, 30 May 2012 (BST)

Try natural=strait. Since the outline of the way is not rendered, I tend to draw a way through ( or a very simple area in ( the feature in question. --Gorm 13:55, 19 June 2012 (BST)

Way should also be allowed

For narrow features like fjords or straits (perhaps natural=fjord and natural=strait), I think a simple line laid roughly down the middle of the feature will do fine. This already works fine both in Nominatim and osmarender. It should be simple to convert the length of this way (area of closed way) -> importance -> rendered label size. --Gorm 14:04, 19 June 2012 (BST)

Agreed. I've added this possibility. --Jeisenbe (talk) 04:24, 21 December 2018 (UTC)

tags for Bay

for objects such fjords probably should use bay=fjord etc. and name=Sognefjord :dr&mx 16:37, 11 September 2012 (BST)

natural=bay == place=bay

This kind of large land formations could better be tagged with place=* For instance the natural=bay+name=German Bight might as well be place=bay+name=German Bight or even can be combined with natural=bay+place=bay+name=German Bight, making natural=bay redundant. --Skippern 17:56, 12 September 2012 (BST)

More thoughts on tagging

The vast majority of bays are tagged with nodes right now. Mappers could use the old fuzzy=* proposal to indicate the scale of the bay.

Another approach for areas: Draw ways that represent only the water bounds of the way (left=not bay, right=bay), which mimics how bays are sometimes defined in reality ("The bay is east of a line drawn from headland X to headland Y."). Again, those lines could be fuzzy=*. Not having to deal with coastline would generally simplify mapping but would likely be more difficult for data consumers, who'd have to deal with a feature defined by a heterogeneous chain of natural=bay--natural=coastline ways. Mrwojo (talk) 00:47, 23 March 2013 (UTC)

Not needed at all:
  • either the ways delimiting the bay are coastlines, in which case they are precise (minus the common imprecision effect of tidals on all coastlines), and the delimitation stroke (or contrasting line between filled areas) that will be drawn is the same as the one used for that coastline.
  • or the ways are not coastlines and they are necessary fuzzy by nature. This means that there is water on both sides, and no stroke should be ever drawn (unless this way is also an administrative boundary or protected area boundary, in which case this is when drawing the admin area or the protected area that you'll get a stroke).
  • Waters on both sides of such fuzzy lines should be filled with identical colors, or alternatively the bay could be given some fuzzy "shining" effect along the costline. If there's no reason to give a different filling color (or semi-transparent shining effect) for the surface of bays, renderers should not fill them at all, leaving the surface of the sea or lake unchanged (this will also avoid filling areas where there are some islands which may have been forgotten as inner members of a relation).
Defining ways as polygons is still better than using as one node, because it explicitly indicates its average extension, and using such a polygon allows more freedom for placing the label naming the bay, and offers a way to discriminate small and large bays at small zoom levels. The only caveat is that you must not forget the coastlines of islands as inner members of a relation (and in such a case, the natural=bay should not be used on a single closed polygon but always on a multipolygon relation.
What this means is that you'll never encounter any conflict between natural=coastline (on ways) and natural=bay (on multipolygon relations). And in fact this tag natural=way should never be used on ways (which should anyway be closed if it ever occurs, and there's no island) but can safely be used on multipolygons (having always at least 2 members: at least 1 coastline member way and at least 1 untagged member way, none of them having this tag). — Verdy_p (talk) 11:15, 29 April 2013 (UTC)
How fuzzy is it? That depends on the bay and even the particular line: some have negligible fuzziness due to distinct landforms while others vary by kilometers. I think it's better to have mappers tag that if they can.
The point of the way tagging idea was to achieve the polygon goal while keeping it simple for coastal bays. Typical cases might then have just a single way across water instead of a multipolygon that additionally maintains a list of all relevant coastline and island ways, which feels redundant and particularly burdensome for gigantic bays.
Just a thought. Mrwojo (talk) 20:55, 29 April 2013 (UTC)

Bays vs Gulfs / Seas

There isn't a clear dividing line between place=sea and natural=bay it the case of large gulfs and seas which are mostly surrounded by land, like the Gulf of Mexico or the Sea of Bothnia vs Bay of Bothnia. Generally there is some size where a bay becomes a gulf or sea, but I don't know of a clear cut-off.

But I believe one clear difference is that bays are an inlet of one water body into one landmass: one continent or island. This means that the "Bay of Bengal" should not be mapped as a bay if it's considered to extend down to the islands of Sri Lanka on the southwest and to be bordered by the Adaman and Nicobar islands on the east. Similarly, [ Baffin Bay between Canada and Greenland is a large body of marine water that is surrounded by several large islands, but not fully enclosed by land on any side, therefore it should also be a place=sea.

(I imagine these marginal seas were named "Bays" by the English rather than "Gulfs" or "Seas" due to the pleasant alliteration ("B_ Bay") rather than any objective criteria? In Indonesian they are both a "Teluk" or "Sea": Teluk Bafin and Teluk Benggala)

While the Gulf of Mexico and other large "Gulfs" could fit the definition for a place=sea or a natural=bay, I would recommend mapping large gulfs as place=sea. --Jeisenbe (talk) 05:42, 28 April 2019 (UTC)

Parts of bay may be land?

The lede discusses that parts of a natural=bay can be land, but my understanding is that Carto and other data consumers interpret this tag to be water. If that's true, then this tag really isn't used on land areas and that description needed to be changed. --ZeLonewolf (talk) 17:39, 10 March 2021 (UTC)

Carto is not interpreting it as water (no area fill), but is interpreted as water-related landform (blue label) Mateusz Konieczny (talk) 21:18, 10 March 2021 (UTC)
Ah, ok thanks. The rendering is a bit misleading, as it shows the blue background but that's not actually part of the render of this feature. Not quite sure how to do better though. --ZeLonewolf (talk) 21:25, 10 March 2021 (UTC)