[RFC] Feature Proposal - Flowlines

[Minh_Nguyen (Minh Nguyễn)]

@dieterdreist

does it? Afaik in a lake there is no river flow, there are waterways entering and leaving the lake, but they do not continue inside the lake. These rivers inside of lakes are just arbitrary flowlines

Flowlines have arbitrary geometry, but rivers can also run through lakes with nonarbitrary geometry. Where a river is dammed, the flowline can be distinct from the thalweg, which can still be mapped as a waterway=river as an exception. In principle, if both are mapped, data consumers would use one or the other depending on the use case. As someone without much experience in hydrology, I also note some seeming inconsistency between the discussions on this forum and the documentation, which focuses on the flow rate but not the method for determining the geometry.

[dieterdreist (Martin Koppenhöfer)]

@Minh_Nguyen

Flowlines have arbitrary geometry, but rivers can also run through lakes with nonarbitrary geometry. Where a river is dammed, the flowline can be distinct from the thalweg, which can still be mapped as a waterway=river as an exception. In principle, if both are mapped, data consumers would use one or the other depending on the use case.

Here’s a visualization of actual flow (at a specific date) in the lake of Constance (biggest lake in Germany), here it is wind induced, which is the main factor for flow in this lake (i.e. it changes), that’s far from what we have in OSM:

Arbeitsgebiete

also, there is no “river” here: Relation: ‪Rhine‬ (‪123924‬) | OpenStreetMap

On this website, they also have a description (in German) how the inflowing river water behaves in the lake.

[Minh_Nguyen (Minh Nguyễn)]

Yes, Lake Constance is an example where any waterway running through the lake would be arbitrary and counterintuitive. There would be little use for the precise geometry of the waterway except in specialized renderings or analyses.

I wouldn’t want to place too much importance on a lake’s water flow. Obviously it matters for identifying a waterway in the first place, but to some extent it’s ephemera, like any other kind of water current. I think it’s important to continue to map rivers where the physical river channel runs through the waterbody – namely, when a river has been dammed, creating a reservoir around the river channel, which remains. The water flow is not relevant to mapping this feature, but I wouldn’t be opposed to also mapping a flow line if desired.

Maps, especially topographic maps, label the channel as the river, not any synthetic lake centerline or modeled flow line. This is consistent with how laypeople understand the location of features along the lakeshore. It’s also consistent with how we handle a river that has been raised with locks and dams. Technically that creates a reservoir too, but we don’t have a separate name for it.