Going back to my original post on this topic, I was looking at OSM recently for my area and noted that some Vicmap Hydro “Watercourse” lines (feature_type_code = connector structure) have been added by others.
From local knowledge, I know that some of these lines are in fact storm-water pipes that outfall into Port Phillip (Bay). An example is the tunnel = culvert & waterway = stream Way: 1444884217.
Can such features be amended to man_made = pipeline and substance = rainwater where I know this is the case? (checking against the local council website at: https://mornpen.spatial.t1cloud.com/spatial/IntraMaps/?configId=a876603d-3827-4328-b2bc-bad1485f65d7, choose ‘Assets’ then tick the ‘Pipes’ layer)
Thanks, John
If they span more than the road width I would agree with your suggestion. Culverts than run for 100s of metres just look weird to me.
It may be worth keeping the waterway tag with flowline as this would show the logical connection between the upstream waterways and the bay.
Do we have the OK to use their data though?
Graeme, I’ve compared what’s on OSM for Vicmap Hydro - Watercourse data against the Victorian Digital Twin (Digital Twin Victoria) as well as the local council mapping website. There is also my local knowledge for where the stormwater drainage outfalls are, as most days I do a morning walk along the foreshore reserve for exercise.
I can always chase up a contact or two at council about permission if required.
Regards, John
1 Like
According to Key:waterway - OpenStreetMap Wiki man_made=pipeline is for “pipe flow”, so the tag wouldn’t apply for these storm-water pipes that outfall into the bay. A man_made=pipeline would be for the pressurised and pumped freshwater supply, or natural gas pipelines, etc.
According to the wiki it would still be waterway=drain.
You could add location=underground to say it’s buried.
Ideally we’d have better ways to tag differences like buried concrete drainage pipes, excavated tunnels, concrete lined canal that is either open to the air, or covered over…
+1 to Andrew’s suggestion about culverts being limited to crossing a roadway.
Because their data is not licensed in a way that can be used in OSM, it’s best not to check against it at all, but instead map based on sources we can use like Vicmap Hydro and your local knowledge.
I just love the definition of pipe_flow and free_flow. Clearly written by someone with good itention, but not enough knowledge. If I came across this at work recently, I would ask if it had been written with AI.
OSMese aside, as it stands, no water pipe should be be in the pipe_flow category as air is allowed into (and out of) the closed conduit by design. Else they could break.
A two part question with a one part answer to identify free_flow or pipe_flow doesn’t help. ‘yes’ to what?
Sewers are problematic too. Closed pipes, usually mostly filled with air. Some under pressure, some vacuum. May carry sewage, surface (storm) water, and often both.
Pipeline goes on to say it is for ‘major’ pipes, but does not define major.
I guess most mappers, don’t want to be tag developers, they just want to easily select the correct tag so they can map the thing they are wanting to map.
But tag development, including the tags and values, their documentation on the wiki, editor presets, fields and quests, local mapping guidelines and tagging examples is just as important, or even more so as the actual mapping work.
Even then, much of development of our tags is likely by non domain experts, to generalise for the most part, we aren’t civil engineers, or traffic engineers, or locomotive engineers but we try to map out all the technical aspects of the power network, the sewage network, even the railway signals train drivers rely on.
Please consider helping out to improve the tags, especially if you’re a domain expert!
as a non-expert, I assume the distinction is that man_made=pipeline consists of gas or liquid that is either pumped through the pipe, or if by gravity then from a large tank (like a drinking water tower) that rarely is let empty, like emptying the water out of the bath (like from the town water tower to your kitchen tap). A stormwater pipe, being empty when it’s not raining would therefore just be a drain, even if made up of concrete pipe assemblies, since it’s the function that defines man_made=pipeline not the form.
Looking at the wiki, it seems then that for a stormwater drain, better tags would be waterway = drain, tunnel = flooded, location = underground & substance = rainwater?
If it’s actually a tunnel, and not an open air drain then yes.
operator=* and operator:wikidata=* to distinguish between Melbourne Water vs local council.
No, I think the opposite is the case. Does a highway stop being a highway if there is no traffic? A stormwater pipe is still a stormwater pipe even if it is empty. The challenge of deciding to use one pipe for another contents is significant, but has been done.
Bit of CV. I started recording UK water mains using pen on paper. GIS looked interesting around UK privatisation, the first GIS (GFIS) I used did not have a mass use viewer, Unix editors only… I’ve converted data models and combined databases. Desktop GIS and SQL server. Set up full network analyses for business planning purposes (not altogether useful given how the UK water industry is regulated to be cheap and nasty.)
However I think none of that is of any use in OSM. There is just far too much resistance to change and the use of automation to correct errors (my impressions from the last 2 years or so reading proposals here). Proposing new tags is hard enough, changing them seems to be nigh on impossible. Sometimes for good reasons (entrenched use), but I think there is a fear of making one piece of good data into bad data in automation that overrides making 99 pieces of bad data into good. While I would agree that there has to be a line drawn somewhere, an extreme bound seems the norm.
I have generally kept away from making water network changes. I’m aware of those local to me that are wrong, but I no longer have access to the correct(er) data, which would not have been licenced for use here anyway. You can’t just join manholes and assume you have the right connectivity…
open_channel or closed_pipe would be better physical feature class definitions IMO. open_channel has a free surface (and can flood and have other interesting hydraulic effects), but a partially filled sewer pipe also has a free surface, so is not closed pipe flow in an engineering sense. I’d leave out the issue of air in the conduit. I’d leave the flow type out of any definition and leave that to network analysis of how particular feature are used, which we may or may not know even if we can see the feature. A pipe bridge could be potable water under pressure, it could be waste water at half depth by gravity or under pressure (pumped).
All this is academic here. I think the likelihood of such a proposed change being accepted is up there with me going into space.
Yes, but we don’t have any tags documented for a “stormwater pipe”…
Sure but there are many more possibilities than just those two.