|
Long ago, a river was dammed up to fill a cleared-out area, which became a lake or pond. The river still flows into the lake, and out of the lake (at the dam). The lake and river have their own names. Should the river be: (1) mapped as a continuous single way into, through, and out of the lake, even though there is no well-defined path of the river within the lake; or (2) mapped as a way that flows into the lake and ends at the edge of the lake, and another way that starts at the dam and continues downstream, perhaps with the 2 ways in a waterway relation? |
|
IMHO, there should be a line drawn through the water body to show the direction of flow. I assume in this case that the river has the same name upstream and downstream of the lake/reservoir/dam, so there should be a connection line tagged waterway=river and probably have the name=xxx River tag too. I notice in large lake/reservoir/dams that the creeks that flow into the waterbodies mostly don't have flow lines through the waterbodies so I assume that generally only the major waterways are drawn through the waterbodies. Example: https://www.openstreetmap.org/way/30324034#map=13/51.1825/8.9752 I notice that the smaller waterways are also drawn as a way though the reservoir and join the main Eder River line. Yes, I should have said that the river has the same name upstream and downstream from the lake. Before the dam (which was maybe 80 years ago) it was just a single river. Thanks for the example. I've done one like that (much smaller), but before doing a second I thought I should ask. Besides the fact that it renders oddly (river name appears inside the lake) - I know, we don't map for renderers - I still am not happy with making up a path for the river within the lake. Because what really goes on there is that the river flows into the lake, and the lake flows into the continuation of the river. So shouldn't it be 2 separate waterway=river ways?
(07 Nov '16, 22:47)
ljb_nj
I think the river flows into the lake until it is full and then the overflow continues downstream. In many similar situations the inflow may be continuous and a permanent watercourse, but the outflow may be intermittent. Other tributaries/creeks/drains and rivers may also feed the dam/lake/reservoir and at times may supply all the flow but I think the convention is to continue the main river system through this lake, and any other bigger water bodies that may form in it's course. The less significant flows should join the main flow line in all cases I can think of.
(08 Nov '16, 07:53)
nevw
This question is unanswered (with only one user responding) - I'm also looking for more info from those in the know. I can't find anywhere in the wiki where this question is addressed. Is nevw correct?
(31 Mar '17, 23:34)
GregRetro
There is no right or wrong here. But nevw's explanation sounds reasonable. If not then add your own answer explaining why the river should be broken up according to you :)
(01 Apr '17, 07:25)
scai ♦
It is though somewhat speculative even if acceptable. Having a "river-line" over a lake object indicates a kind of contradiction. At the same time it helps map-makers to accomplish the river brake because a longish lake, crossed alongside by a river-line, between two river areas is probably just the missing river section.
(01 Apr '17, 08:08)
sanser
3
@GregRetro which people do you mean with "from those in the know" ? newv's answer is upvoted 6 times, meaning that at least 7 people think this is the correct way to map it. A page on the wiki does not have more weight than an answer here
(01 Apr '17, 11:25)
escada
As someone working on waterway cartography and navigation, I would back up nevw's answer, not least because a continuous polyline through the lake makes routing much more likely. (It's theoretically possible with a polygon alone, but in practice most commonly used OSM routing software doesn't route well through polygons.)
(01 Apr '17, 18:02)
Richard ♦
@escada There weren't any upvotes on the answer prior to me bumping the question. Thanks all... @Richard Are you talking about using polylines (i.e. water=riverbank) instead of a river "way" through the lake? Or am I confused by your use of the term "polyline"? One more thing - is it feasible to continue the river (as waterway=river) through the lake, but have that portion of the way be unnamed? It would seem to me that routing would still work, and there would be no problem with name label clutter.
(01 Apr '17, 22:13)
GregRetro
BTW avoiding the waterway through a pond/lake raises this error in osmose: http://wiki.openstreetmap.org/wiki/Osmose/issues#1220
(23 Jul '17, 17:38)
juminet
showing 5 of 9
show 4 more comments
|
|
From a cartographic point-of-view, having developed a style myself, I would be strongly in favor of only mapping the main river system flowing through the reservoir as "nevw" pointed out, and seems de-facto convention. Most other tributaries/creeks flowing into a reservoir will be insignificant compared to the main flow. Adding them inside the lake by mapping and connecting them to the main flow with ways, will clutter the rendering and cartographic result, especially if these tributaries have names themselves and get labeled. This will be very confusing in terms of cartographic results. Imagine a single reservoir / lake with dozens of water related name labels inside its extent... 1
This is from the renderer's point of view. I guess data consumers prefer having a river line through the lake. Both views have their pros and cons.
(01 Apr '17, 16:14)
scai ♦
|
|
From a topological network point of view I would certainly create a centreline for any water body where a stream flows into then out of. I would start and stop at the bounday of the lake. This would create a hydrologically connected network that network analysis could be applied to. If you did not then you would get a break in the network and for me this is one of the major limitations of the OSM data. As a side note rivers don't flow up hill! So the line should be digitised in a source to sea direction. |
|
I really don't agree with the answers. When I look the map "openstreetmap standard" for the Leman Lake, I'm very shocked to see in the middle of the lake a repeated name "Le Rhône" : https://www.openstreetmap.org/#map=14/46.4371/6.7584 And when I look at the "lac d'Annecy", it is worst : https://www.openstreetmap.org/#map=14/45.8838/6.1560 I'm living in Annecy and the mean source of the lake is not a river but a spring under the lake. To see a line with the name "l'eau morte" is not very bad, it is completely false. The lake is alimented by the spring under the lake but also by 10 or 12 other small rivers. Why to have choose one of them ? And there is only one river out of the lake, a new river with a new name, no one of the 12 incoming rivers have this name (nor the spring under the lake). Why don't choose the name the outcoming river rather that one of the 12 incoming ? And if you go to Annecy or to Geneva you will see that there is no flow at all in the lakes !!!!! On wiki, it is written for RIVER, that you have to do so (a way in the middle of the river + lines for the banks) because the river is between the banks and flows betwwen them in one direction. To see in the middle of the river (at the way of the main stream) the name of the river is of course the good way because it is a river. A lake is not a river, there is no flow, to put a line in the middle of the lake is bad for the map (line of names) and false because it is not the reality at all. On wiki nowhere it is written to do so for a lake. I think that, in the contrary, someone should write not to do it. Best regards |
