[Imports] UN Mappers import of UNSOS waterways in Somalia
arnekimmig at icloud.com
Thu Apr 9 17:56:24 UTC 2020
Hi Christoph and Rafael,
I agree with Christoph that only somehow active rivers should be mapped as waterways.
I mapped a few (not all) rivers in the import area only where I saw some erosion banks or gullies.
See https://www.openstreetmap.org/#map=13/2.7975/41.1247 <https://www.openstreetmap.org/#map=13/2.7975/41.1247> for some examples and compare them with the rivers to be imported.
At a first glance it seems to me that mapping only rivers with erosions features leads to a much less connected river system than the one represented by the import data. Which one represents reality better?
For practicability and to come to action I would think that we did discuss all important aspects of this river import and I would propose that we should start with the first subproject as a kind of prototype when the workflow is ready. Performing the import and correcting the data we could discuss further and gather new experiences to improve the workflow and the mapping quality for the subsequent parts.
Just one more remark concerning tags: I’m not a friend of inventing new tags like intermittent=decade or so. For questionable rivers I would instead use an [annotation="maybe old riverbed"] or a [fixme="confirm active river?"] or similar to show that local know how is needed.
For the next about 2 weeks I will be off somehow for this project but afterwards I’ll come back.
See you and regards
> Am 09.04.2020 um 18:33 schrieb Christoph Hormann <osm at imagico.de>:
> On Thursday 09 April 2020, Rafael Avila Coya wrote:
>> If a river gets dry for any reason (let say for some decades ) but u
>> could still see the river course (which is only possible in very rare
>> occasions) u could still call it a river just to maintain the
>> landmark notation and geography."
> I have not made a statement on what can be called a river and what not.
> That is a speciality problem of English language semantics. What i
> said is that a feature that shows no verifiable water flow in present
> times is not to be tagged as waterway=*.
>> I am not an expert, and I would do what Arne suggest if I was mapping
>> without discussing this with anybody, but still, like Arne, I see the
>> lack of a tag for these kind of rivers.
> Feel free to introduce a new tag for something verifiable for which no
> tagging exists yet. But you should keep in mind that there are already
> a lot of tagging ideas in that wider domain that have either failed or
> are used without a well defined meaning:
>> As for the Lagh Dara, this colleague has sent me two images of the
>> rivers passing by Afmadow town.
>> One shows the river as we can see it normally:
>> But the second shows how it was during the 2018 devastating
>> floodings: https://flic.kr/p/2iNyA7z
>> He has been on UN missions in Somalia many times, so he has too on
>> the ground knowledge.
>> By the way: the Lagh Dera river geometry is right now of very poor
>> geometry in OSM, while the UNSOS one is very good in comparison:
> Please keep in mind that anecdotal examples are not representative for
> the whole. I could easily provide a dozen examples for the existing
> mapping being fairly accurate and the import data being off by at least
> Flooding is a completely different subject. We do not map 2/3 of
> Bangladesh as water just because it becomes flooded on a regular basis.
> waterway=river/stream is for mapping localized, verifiable surface
> water flow, it might be seasonal, it might be ephemeral - but it needs
> to be there. What your example image shows is specifically the lack of
> water flow over longer distances. Water from heavy local rain does not
> drain on the surface along the supposed river but accumulates in local
> depressions as temporary lakes.
> Overall i think we are departing from the matter of import review here.
> Christoph Hormann
> Imports mailing list
> Imports at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Imports