[Talk-GB] Boundary_line at the coast

[Colin Smale]

On 2019-09-07 23:06, Edward Bainton wrote:

> I'm interested in boundaries marked at Mavis Grind [1] (thanks to SK53 for the waterway=portage [2] tag - Mavis Grind is an old Norse portage, still in use by Shetland Canoe Club). 
> 
> 1. Does anyone know if county boundary lines at the coast are set at mean low water? There's a gap between coastline (which I understand is MHW) and the county boundary: https://osm.org/go/e7tUHoc5F--?m=&way=669235281

Indeed, GB administrative boundaries are set to MLW(S) and the coastline
is defined in OSM as MHW(S). So normally the coastline should be
"inland" relative to the admin boundary, although in areas with steep
cliffs, harbour walls etc they are shown as coincident. 

> 2. "coastline" is very coarse - is it ok to make it follow the coast more finely, or is it some important legal line where it stands? (I've already done this on the east side of the portage, but then thought perhaps that was a no-no: https://osm.org/go/e7tUNAM7G?layers=D&m=&way=669235281)

The coastline is in most cases (source=PGS) derived from very old aerial
imagery. A better source is OS Open Data but it is a lot of work....I
have done bits and pieces around Scotland, but not this bit of Shetland
yet. The actual physical coastline doesn't have any particular
administrative significance, AFAIK. 

> 3. Also, there are two walls visible on aerial imagery that all but match the doglegged county boundary as it crosses the isthmus. Is it safe to assume that these mark the actual boundary, and can I tug the boundary to match them? Or maybe assume the boundary is definitive, and the imagery is misaligned, so I should move the walls? Or leave well alone? 
> West side: https://osm.org/go/e7tUHqdxn?m=&way=669235281 
> East side: https://osm.org/go/e7tUHrpqY?m=&way=669235281

Note that the OS data was mostly surveyed using high-precision GPS
equipment (±1cm or so); its lat/lon positioning is going to be far more
accurate than what we can get from either aerial imagery (±5m
sometimes?) or consumer GPS (can be ±10m or more). So I tend to accept
that Boundary-Line is pretty much correct, and sometimes things have to
move to fit that. In this case, if it can be established that the walls
are supposed to on the boundary, then I would consider moving the walls
to fit the boundary, and not the other way round. 

Having said that, when it comes to high water / low water marks, I
believe OS re-survey the coast every few years, and make a lot of use of
aerial imagery for this. Don't forget when looking at Bing etc that you
don't know what the state of the tide was at that time, and with a
shallow slope or sandbank it might be difficult to say with any
certainty where the water stops as even sea water is fairly transparent
through short distances. 
  

Links:
------
[1] https://osm.org/go/e7tUHqY_g-
[2] https://wiki.openstreetmap.org/wiki/Tag:waterway%3Dportage
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20190908/2eac7715/attachment.html>