The end of waterway map

SomeoneElse · 2024-01-29T23:17:29.617Z

Any idea what it’s detecting at https://waterwaymap.org/ends/#map=22/53.95422459/-1.0778698? Is it confused by the lock or does it not think a river should flow into a canal?

sbre · 2024-01-29T23:51:31.482Z

I see a lot of small black dots on low z-levels, which disappear when I zoom in. Since the data for smaller waterways apparently then exists, I would like to be able to view them to fix issues in my local area.

amapanda_ᚐᚋᚐᚅᚇᚐ · 2024-01-30T10:07:28.497Z

SomeoneElse:

Any idea what it’s detecting at https://waterwaymap.org/ends/#map=22/53.95422459/-1.0778698 ? Is it confused by the lock or does it not think a river should flow into a canal?

That’s been fixed now. Initially I wasn’t including waterway=canal. But I kept seeing issues like that. So now I include a waterway=canal iff there is lock=yes. Water will now flow through locks.

I made this code change yesterday, and posted this message before that code change. This also fixed one of the erorrs I mention in my original post. Ooops.

SekeRob · 2024-01-30T10:40:04.098Z

Coincidentally had a flag recently, then saw the end node was marked with something which translates to a sinkhole (a ponor). Remembered that as we have many little streams that end in nothing.

waldyrious · 2024-01-30T10:40:24.508Z

I found an interesting case, where a water course is mapped as diverging into an underground/overground pipeline which then routes onto another water course: WaterwayMap.org - End points.

The pipes are mapped as Way: 151069423 | OpenStreetMap, Way: ‪Conduta de água‬ (‪225455540‬) | OpenStreetMap and Way: 225455539 | OpenStreetMap, and their direction is correct. Shouldn’t the start of the pipeline not be considered an end for the water course?

Edit: here’s another one that I’m a bit puzzled by (also involving pipelines, this time one that diverges and returns to the same water course).

mikeocool · 2024-03-08T22:17:31.909Z

The title of this post made me very sad, the content made me very happy. Thanks for a great map + tool!

amapanda_ᚐᚋᚐᚅᚇᚐ · 2024-06-06T12:34:35.344Z

waldyrious:

The pipes are mapped as Way: 151069423 | OpenStreetMap

w151069423v11 is a usage=penstock, an underground pipe transporting water. However it’s tagged waterway=canal, so it’s excluded from this calculation. The wiki says to use waterway=pressuried for that.

Do you think it’s really a =canal? We can change the WWM tag filtering rules id needed.

waldyrious:

Way: ‪Conduta de água‬ (‪225455540‬) | OpenStreetMap Way: 225455539 | OpenStreetMap

These 2 don’t have a waterway tag at all! . Is the first really called “Conduta de água”? Remember, name is not for descriptions.

amapanda_ᚐᚋᚐᚅᚇᚐ · 2024-06-06T12:55:10.091Z

`osm-river-basins`: Website to show how are rivers in OSM connected

It managed to find an end of 20,000 km in my neighborhood that I can’t fix – it’s an old riverbed connected to the main flow by an artificial canal with a pumping station.

(orig. post)

I’ve long wondered about half the Danube being emptied into a field in Serbia, I’m glad to see we have @Duja to manually survey!

The problem is n6731666597. Water flows in from w716661430 (which is tagged waterway=river), and only flows out through w716661432, which is only waterway=canal. After 2 (a & b) other =canal ways, it rejoins the Danube at n10098027262.

I can’t see any mapped pumping station in OSM . Are you sure =canal is correct for w716661432? Does the water flow along that way? I thought canals were “always” still water. If you change the tags to something else which is more accurate, it would make the calculations work. waterway=canal,flowing_water=yes sounds too much like (mis)tagging for the ~~renderer~~ waterwaymap, so I don’t want to encourage it.

ⓐⓜⓐⓝⓓⓐ

Duja · 2024-06-06T13:34:27.993Z

Actually, I was the one who originally mapped those waterways…

Mali Begej is actually an oxbow lake, i.e. the water is mostly still. It collects a lot of rainwater from surrounding ditches. However, I had no better option than a waterway=river, since waterway=oxbow is not a thing (and would not render).

amapanda_ᚐᚋᚐᚅᚇᚐ:

I can’t see any mapped pumping station in OSM .

There’s one in your point a, as well as one at the northern end of Mali Begej (Kerektov). Those regulate water levels in the Mali Begej system by pumping water to or from the Tisza (not Danube!). That’s a common setup, there are several similar oxbow lakes in lower Tisza.

amapanda_ᚐᚋᚐᚅᚇᚐ:

Are you sure =canal is correct for w716661432? Does the water flow along that way? I thought canals were “always” still water.

Well, it is still… but the “river” is also still. But we also have proper, “flowing” rivers which flow into canals, (Relation: ‪Karaš‬ (‪9811356‬) | OpenStreetMap ; Way: ‪Стари Бегеј‬ (‪958660502‬) | OpenStreetMap). This is a marshy area, all rivers are really slow, and many canals were dug in the last two centuries to regulate waterways.

Simply, I don’t think you should presume that canals are still water in the same way as seas or lakes. Many of them actually carry significant volumes of water from incoming natural or artificial waterways. But even if we accept that premise, something is odd with your algorithm – it seems that those 20,000 km are the complete upstream headwaters of the Tisza…

I suspect the culprit is this piece of pipeline tagged waterway=river, which I’m willing to sacrifice to the renderer, and retag it as a pipeline (which is more apt anyway).

waldyrious · 2024-06-06T13:46:40.615Z

amapanda_ᚐᚋᚐᚅᚇᚐ:

waldyrious:

The pipes are mapped as Way: 151069423 | OpenStreetMap

w151069423v11 is a usage=penstock, an underground pipe transporting water. However it’s tagged waterway=canal, so it’s excluded from this calculation. The wiki says to use waterway=pressuried for that.

Do you think it’s really a =canal? We can change the WWM tag filtering rules id needed.

Oh, it definitely seems like an instance of mistagging. I’ll change waterway=canal to waterway=pressurised. Thanks for the hint!

amapanda_ᚐᚋᚐᚅᚇᚐ:

waldyrious:

Way: ‪Conduta de água‬ (‪225455540‬) and Way: 225455539

These 2 don’t have a waterway tag at all! .

I assumed that “Waterways (inc. canals, etc)” included other water circulation infrastructure like (in both of these cases) man_made=pipeline + substance=water, rather than strictly ways tagged waterway=*. What should be done in these cases? Should they also be changed to waterway=pressurised + usage=penstock?

amapanda_ᚐᚋᚐᚅᚇᚐ:

Is the first really called “Conduta de água”? Remember, name is not for descriptions.

You’re right, that’s a tagging mistake. I’ll remove the name.

Duja · 2024-06-07T09:58:44.531Z

Duja:

I suspect the culprit is this piece of pipeline tagged waterway=river, which I’m willing to sacrifice to the renderer, and retag it as a pipeline (which is more apt anyway).

Alright, I’ve just deleted (OSMCha ) the short piece of “virtual” waterway=river that was feeding into that pipeline, and tagged the latter as waterway=pressurised (and did the same at a few similar locations around). That should separate the whole Mali Begej system from the Tisza, and let the Tisza on its merry way into the Danube. When the waterway map refreshes, it should show an endpoint of ~200 km rather than 20,000 km; that’s arguably correct, since the Mali Begej constitutes a separate hydraulic system, so the water of all the infeeding drains must be pumped into the Tisza.

Similarly, in @waldyrious 's examples, “virtual” waterway=rivers that feed into pipelines should be somehow eliminated…

However, as we recently discussed in Should river lines be mapped through lakes… , those “virtual” waterways that feed from rivers into pipelines are beneficial for the waterway topology analysis, so I’m uneasy with their immediate deletion. I’m inclined to just connect the pressurized ways to the main waterway, since we don’t know where’s the underwater end of that pipeline anyway (and it’s all just an approximation).

waldyrious · 2024-06-07T11:14:41.415Z

Duja:

Similarly, in @waldyrious 's examples, “virtual” waterway=rivers that feed into pipelines should be somehow eliminated…

However, as we recently discussed in Should river lines be mapped through lakes… , those “virtual” waterways that feed from rivers into pipelines are beneficial for the waterway topology analysis, so I’m uneasy with their immediate deletion.

In some cases I’ve been using waterway=link to complete the topology. In other cases I’ve been using relations with type=waterway + waterway=river, and tagging the individual members with role main_stream or side_stream, accordingly. Either approach seems preferable to deletion. What do you think?

Duja · 2024-06-07T11:28:37.648Z

I was unaware of waterway=link, it must be a recent one… yes, it was added to Wiki only in 2023, and we touched upon it in the linked discussion. Sounds like a good solution, and I intend to embrace it.

(I would also like that we officially introduce waterway=oxbow , currently with mere ~200 uses worldwide (overpass turbo), but I was always lazy to formulate and put forward a proposal. Guess I’ll just document the existing uses for the beginning. We have documented water=oxbow but that one is inferior – it’s more natural that they are modeled as waterways first).

amapanda_ᚐᚋᚐᚅᚇᚐ · 2024-06-10T09:14:24.217Z

waldyrious:

In some cases I’ve been using waterway=link to complete the topology

I think waterway=flowline is better than =link. (the waterway=link creator @quincylvania agrees).

waldyrious · 2024-06-10T14:00:49.363Z

Thanks for the heads-up! I’ve edited waterway=link and waterway=flowline to mention each other explicitly, and clarify their intended use.

I think it might be good to locate waterway=link ways that are connected to water features on both ends, since they might be instances where people actually meant to use waterway=flowline. Is something like this possible with Overpass, or dos it require more advanced querying systems?

In the meantime, I queried for all waterway=link elements in Portugal (the area I’ve been working on based on WaterwayMap for the past weeks) and indeed most if not all of them were created by me. I’ve edited them to use waterway=flowline instead.

(edit: I went ahead and searched globally —it seems like if we exclude ways with canoe=*, the query returns mostly ways that were meant as flowlines— and edited a bunch of them as well: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13. /cc @RedAuburn who did a few of these in the UK, @dknelson9876 in Utah, @CJJ2501 in South Africa and @Swedneck in Sweden.)

amapanda_ᚐᚋᚐᚅᚇᚐ · 2024-06-10T14:55:42.196Z

Duja:

I had no better option than a waterway=river, since waterway=oxbow is not a thing (and would not render).

you could map the area as natural=water, and water=something.

Duja:

the Tisza (not Danube!).

oops!

Duja:

something is odd with your algorithm – it seems that those 20,000 km are the complete upstream headwaters of the Tisza…

Yes and no. When a waterway split in 2, the total upstream length is shared equally. So this 20,000 km can include a lot of the Danube’s upstream, Sometimes smaller rivers like this can be responsible for big upstreams.

Duja:

This is a marshy area, all rivers are really slow, and many canals were dug in the last two centuries to regulate waterways

I live beside the Rhine, which was also straightened out in the 19th century. I’m no stranger to paths being flooded, or oxbow lakes. While cycling this weekend, I also saw some under construction pumping stations to regulate water flow. (We’ve currently got some high water around here recently)

Duja:

Simply, I don’t think you should presume that canals are still water in the same way as seas or lakes

Hmmm I’d like to try running the algorithm including waterway=canal. I wonder how much has been mapped presuming the direction doesn’t matter.

Duja:

I suspect the culprit is this piece of pipeline tagged waterway=river,

Yeah, if it’s in a pipe like that, it sounds like a river.

Hungerburg · 2024-06-10T23:11:04.145Z

amapanda_ᚐᚋᚐᚅᚇᚐ:

I live beside the Rhine

Close to Strasbourg to Mulhouse Canals split natural waterways –WaterwayMap.org - OSM River Basins

Duja · 2024-06-11T08:57:25.537Z

amapanda_ᚐᚋᚐᚅᚇᚐ:

you could map the area as natural=water, and water=something.

In theory yes, but this is not a natural way of mapping a feature which is ~5m wide and ~30km long. Besides, those waterways often have tributaries and outlets. Just as with rivers, I’d prefer having an underlying waterway, and only an optional natural=water area surrounding it.

amapanda_ᚐᚋᚐᚅᚇᚐ:

Hmmm I’d like to try running the algorithm including waterway=canal. I wonder how much has been mapped presuming the direction doesn’t matter.

I wonder that too; some canals do have a “natural” direction, and with others it may not be so obvious. I also try to point ditches and drains in a “proper” direction, but that is not always obvious or easy (often, there are grid-like networks of ditches where water barely flows).

Besides, I’m having second thoughts about my proposal; having endpoints at canals may clarify where sub-areas exist. Perhaps, as with the main waterway map, you could provide an option how to treat artificial waterways (as end points or not).

Furthermore: many (but not all) canals are separated from the adjacent waterways either by lock gates or by pumping stations. Those sections should be mapped as waterway=lock_gate and waterway=pressurised respectively, and those may represent “proper” end points of a drainage.

dieterdreist · 2024-06-11T13:51:14.000Z

Furthermore: many (but not all) canals are separated from the adjacent waterways either by lock gates or by pumping stations.

or ~~elevators~~ boat lifts.

3000×2000 1.85 MB

OpenStreetMap

Way: ‪Oder-Havel-Kanal‬ (‪25824407‬) | OpenStreetMap

OpenStreetMap is a map of the world, created by people like you and free to use under an open license.

https://wiki.openstreetmap.org/wiki/Tag:waterway=boat%20lift?uselang=en

waldyrious · 2024-06-19T10:27:40.953Z

Just a quick note to mention that I opened PR #10283 in the iD repository to add directional rendering for lines tagged with waterway=flowline.