Talk:Tag:waterway=stream end

Misleading tag?

I don’t recall a proposal for this tag, to me it seems a bit strange to speak about a stream end for a place where the waterway seeps into the ground, because we are also mapping underground streams. —Dieterdreist (talk) 13:46, 12 August 2019 (UTC)

I will prefer waterway=soakhole, because is semantically clear. Something B (talk) 17:49, 1 October 2023 (UTC)

Rivers and other waterways?

Why is this called stream_end, is it limited to streams? —Dieterdreist (talk) 13:48, 12 August 2019 (UTC)

I know of numerous streams whose water seeps into a scree or glacial moraine, without a visible sinkhole or ponor. I assume larger waterways needs a natural=sinkhole for that. Ends of artificial waterways is a different thing. Otherwise, stream_end doesn't sound very scientific, does anyone know the proper technical term? --BáthoryPéter (talk) 15:42, 29 March 2023 (UTC)

Useful and much needed

Streams that end without a connection to another waterway are common in arid climates. The flow of these streams terminates because they lose water to evaporation and absorption. Contrary to some suggestions, these streams do not end in sinkholes and they do not continue underground (except as part of the underground flow of a larger aquifer, which is an entirely different concept).

Sinkholes, in contrast, occur in karst terrain where the bedrock can be dissolved by water to form tunnels and caverns into which water can flow. A sinkhole can potentially divert an entire river underground, and there are cases where this happens (e.g., the Lost River in Indiana). However, the reduction of waterways by absorption and evaporation is much more common. Such as, for example, the various small washes in the Mojave Desert where flowing water evaporates or sinks into the underlying water table.

Before a river could come to a complete end by evaporation and absorption, it would necessarily be reduced to the size of a stream. So, the waterway=stream_end tag is appropriately named.

--B1tw153 (talk) 19:39, 6 July 2023 (UTC)

As you describe it, it should be dedicated to infiltration in appropriate ground. It is not sinkhole, it is not man made works. Then stream_end got a double issue:

* it should not be focused on streams (it is also suitable for rivers)

* it deserves a more particular name than end.

Back in 2018 during this proposal, it was clear OSM was missing the third water flowing regime dedicated to infiltration beside free and pipe flows. Fanfouer (talk) 14:13, 25 September 2023 (UTC)

Proposal:Hydropower_water_supplies was well documented and it starts to touch on some of the cases of natural water flowing below ground.

The case of a river coming to an "end" is somewhat different from a stream that ends because of the volume of water involved.

* A river can end by joining another river, or at a lake, wetland, or other body of water, especially where the lake is endorheic.

* A river could "end" as a surface waterway by disappearing into a sinkhole. For example, the Danube Sinkhole captures the water of the Danube River, which then flows through caverns and emerges as the Radolfzeller Aach. This is not exactly an "end" to the river, since the flow of water underground could be mapped to the extent that it is verifiable. And we have natural=sinkhole to map this case.

* A river can flow primarily underground and partially emerge where the bedrock is impermeable and the bedrock channel is narrower. The Mojave River is an example of this, where the river emerges several times between Victorville and Soda Lake. This can be seen in the aerial imagery of the river, for example, at the Mojave Narrows in Victorville, just north where the river continues below ground, and further downstream where the river emerges again. The channel of the Mojave River carries surface water when the river floods, so it is mapped as waterway=river + intermittent=yes. Eventually, the river flows to Soda Lake or Silver Lake, endorheic lakes where the water is lost to absorption and evaporation. The intermediate points where the river emerges or disappears depend on the water level in the river, and so may not be persistent enough to be mapped in OSM.

* A river can end by forming a delta in an Endorheic basin, such as the Okavango Delta. In this case, the river is fragmented into wetlands and smaller streams, where the water is eventually lost to absorption and evaporation (or some eventually makes its way to Lake Ngami). The wetlands serve to terminate some of the waterway fragments, but a stream fragment that ends outside of a wetland could be tagged with waterway=stream_end.

I don't know of a case where the "end" of a river would be different from these cases and require additional tagging, but if there are such cases I would love to hear about them! Streams, on the other hand, are small enough to come to an end without any other notable features where they terminate.

As for the value of stream_end, I agree that it is not so poetic. Personally, I would be happy with any tagging that captures the concept in a way that fits well with the geology and hydrology and that is broadly accepted by the OSM community. For that, the proposed waterway=stream_end tag seems good enough.

--B1tw153 (talk) 21:27, 26 September 2023 (UTC)

Actual use

Since there seems to be renewed interest in this here a summary on how this tag is practically used (roughly in order of volume of use)

• declarative locations of stream ends (a - usually intermittent - stream verifiably exists and is mapped at some point but verifiably does not extend much further downhill - hence the mapper declares an end location and tags it)
• from imports (mostly Canada) with unknown semantics
• mechanical addition at beginning and end of the extent over which a stream has been mapped remotely (mostly organized editing activities, mostly Africa)
• synonym for fixme=continue on waterways (often combined with a note tag)
• actual localizable waterway ends (natural waterways in certain geologic settings, deliberate percolation locations of small rainwater drainage channels etc.)

The first, fourth and fifth are actually meaningful information - but of little practical use if they are indiscriminately tagged with the same tag of course. Imagico (talk) 14:52, 26 June 2026 (UTC)

We also got inlet=* and natural=sinkhole since to set more precise information. I'm not so keen on this value of waterway=* Fanfouer (talk) 15:01, 26 June 2026 (UTC)