The water is raised using Archimedes screw pumps. My question is about the waterflow after lifting the water: it flows through horizontal concrete ducts or pipes, not open like a drain, and containing water and air.
There obviously is some pressure inside, because of the pumping, but the flow is also maintained by gravity forces. This does not seem to fit in the waterway schedule based on a free flow vs pressurized distinction.
Is it a pipeline? Is it a drain? Is it a tunnel, is it a culvert? Or do we have a specific tag for this object?
Here is an overview example, a different pumping station (three water levels):
Interesting question. Going through the values for waterway=* I came across this one (undocumented?), this might be an option to generally describe the whole section (screw section + duct-sction).
Often we as mappers can not see the distinction between the two (and wether it is a screw or a pressurized pump) because they are often covered. This at least tells you this in no normal waterway, but overall water goes up instead of down and it is not navigable.
Further refinement for the two sections is of course still possible with additional tags, but I don’t know what they might be
I’ve mapped a lot of pumping stations, and I simply tag those with waterway=canal layer=-1 tunnel=culvert
I’m more irked by the (apparent?) lack of tag for a pumping station as a waterway feature. The one at the top of that “culvert” is a man_made=pumping_station building=service , but it does not share any node with the waterway, so any waterway data consumer must resort to spatial analysis (and a hope that mappers did the right thing). Contrast with detailed mapping instructions at waterway=lock_gate.
There’s no reason why a single pumping station should not have pumps within that have different duties, both source and/or destination. So no direct connectivity. Yet if the building is the object used to join the flows, connecmay be created that do not exist in reality.
In pedantic work mode, I used to say “Pumping stations don’t pump, pumps pump.”
And how I am, as a mapper, supposed to know the layout of pumps and pipes enclosed within a locked building, often fenced and on a remote location? Why shouldn’t a point on a waterway suffice to denote that the water is somehow being pumped in whatever direction?
The map is a model of the real world, with any necessary abstractions. In non-pedantic work mode, it’s pumping stations that pump, since they’re immediately verifiable.
(I acknowledge that man_made=pump exists, but it has mere 2,000 uses worldwide, and the Wiki sorely lacks instructions where to put one so as to have useful semantic role in the world).
Very true. So by mapping the connectivity by pumping station, you’ve made an assumption that cannot be easily verified. I would make similar comments about assuming sewer connectivity by manhole covers. In most cases the assumption will be right, but fails on the verification test.
In this particular example, the source and destination can be visually verified, it’s pretty much in the nature of the type of pump. So why not map it as man_made=pump? Twice?* The building could still be a pumping station, although is the pump inside or outside the roofed structure?
*It’s probably a pump set where one pump does most the work. The other might be a standby to do the same job, or might also have an assist role. Possibly verifiable if you spent enough time recording how they operated. I’d map two pumps.
I recall a pumping station with two sets of pumps, one pumped raw water from a source pipe through treatment, another set pumped treated water to supply. The water flow connectivity is through the treatment process of tanks, pipes, channels and filters, not the pumping station. This would probably get mapped as a loop in OSM, which I think would generate a validation error.
But what’s wrong with casually putting a =pump inside the =pumping_station if “The map is a model of the real world, with any necessary abstractions.”?
It’s to keep the water level in the polders low enough. A polder is basically a ginormous bath tub below sea level. Without these pumps, the bath tubs would fill up and we would drown.
(My street is 6 meters below sea level. It is in a polder protected from the sea and the rivers by dykes up to 15 m high.)
This illustrates it quite nicely I think. The polder land is not only below sea level and below the level of rivers, but also below the typical navigable canals. So water is not the lowest point in the landscape, but the highest. The canals are basicly a series af aquaducts between the different polders. In the Netherland we have some 5.000 pumping stations to keep dry feet.
As for the pumping device, I am knowingly ignorant, I just think it’s a node tagged by someone who knows their stuff.
As for hydraulic connectivity vs navigation connectivity, I think both entry and exit are attached to the pump node to indicate hydraulic connectivity, and to prevent navigation we can add e.g. a boat=no tag to the end sections of the waterways. For canoe connectivity portage=yes (I think) is available.
My question was not about the pump (device) or connectivity, just about the physical water exit, between the device and the receiving water body. I think we can agree that it is a waterway.
Is it pressurized or free flowing? Most devices are open to air and raise the water to the exit level; from there gravity does its thing. The water exit is level or slightly downwards.
So it’s definitely a non-natural waterway, but not a pipeline or pressurized.
I am inclined to tag it as a waterway=drain, draining water from the pumping station.
Is it a tunnel? I think yes, just as a culvert is a tunnel. Is it a culvert? I’m not sure. I have heard the opinion that a culvert is a buried free flow connection between water bodies at the same level, not an intermittent exit conduit of a pumping station, and data users should be able to trust that there is free flow water connectivity there. I tend to agree with that.
In the end, I think I’ll go with just waterway=drain, tunnel=yes.
I was actually thinking about what’s inside the building= or location=underground for other cases in general , and forgot about this post’s screw pumps inside the whole man_made=pumping_station
Screw pumps could simply get eg pumped=yes (forced “flow” overlaps with pipe/closed vs open channel flow)
