|
Looking at this national park Puinawai Natural Reserve you can clearly see the river is intended to run around the boundary of the park (i.e. the river is meant to be the boundary/part of the boundary), however they are off by a decent amount. Is this an error in the data or is there some explanation I'm missing? Thanks |
|
The source for this boundary says "SIGOT DANE UNODC-SIMCI OCHA-COLOMBIA" so it's probably part of either an import or drawn from a map. In Either case it was not properly aligned with the other border ( I didn't check which one is the oldest in OSM). It's to hard tell which of the 2 sources is correct by just looking at the map. Properly aligned aerial imagery might help Cheers, I'll try find a reliable aerial image source and maybe adjust the boundary in JOSM before I use the data. It isn't that big of an issue however as the national park is just there as a landmark.
(21 Oct '14, 02:55)
chillNZ
1
Be careful about changing this boundary. Moving it to follow the current river's route may actually be changing correct data to incorrect data. I see two possible scenarios for this boundary: 1. It could have been defined years ago when the river had a different route, and the officially-defined boundary is static and remains unchanged even if the river changes its route. In this case, changing the boundary in OSM would be incorrect. 2. It could have been defined to follow the river wherever its course may deviate to. In this case, changing the boundary to follow the river would be correct.
(21 Oct '14, 18:34)
alester
1
Btw I'm not uploading my changes to the actual OSM database just so it's clear I'm not screwing with the actual source. It would just be for my local mapping application, which is being used in a small live system eventually. I decided just to leave the data as-is in the end anyway, as it doesn't look horrible and I doubt most people would even notice while navigating the map.
(21 Oct '14, 21:56)
chillNZ
Fair enough. Thanks for the clarification.
(22 Oct '14, 00:16)
alester
|
|
Looking at the border and the river I am sure it isn't explained by the river moving in this case. How can it be realigned? If some fixed points could be recorded on the ground with a GPS that could be helpful, but what are fixed points? the villages may be, only some local knowledge could confirm that, do traders visit? with a GPS they could also trace the river, ideally at time close to when the Ariel images are updated. I looked at the area with the Potlatch2 editor which shows how remote it is, the villages are only a few huts, that may get moved. Perhaps there are some airstrips that someone could get excellent positions of. Who flies there? would they give some traces? This is quite a difficult job without some accurate fixes. |
Rivers do change course, so although the river may have been the border at one time it may not be true at this time.
here an example http://socalgis.org/2014/03/12/time-lapse-of-a-river-changing-course/
Hmm interesting, I hadn't considered that the river could have moved. I think I'll just leave the data alone and trust OSM is accurate - it's not really that big of a deal if it's not, just makes my map look slightly inaccurate.