[Talk-us] LAST CALL: Looking to start on reconciling New York City admin boundaries

[Kevin Kenny]

I've asked about this in the past, but not heard much response.

It appears to me that I have only a few days' more work to finish out the
boundaries of Westchester County municipalities, which means that I'll be
starting on New York City.

If anyone has a serious objection to deconflating borough boundaries from
The Coastline, please speak up now!

The conflation of the boundaries with the coastline is not quite correct:
for instance, the Manhattan/Queens and Manhattan/Brooklyn boundaries follow
the east shore of the East River as it stood in 1898.  Landfill and
dredging has moved the shoreline, especially on the Brooklyn shore between
the Navy Yard and Red Hook, but according to both the NYC and NYS GIS
systems, has not moved the formal boundary between the boroughs.  NYC's GIS
(for boroughs) and NYS's GIS (for counties) agree on where the line goes.
The Borough of the Bronx extends to the legal center of the Hudson River
north of Spuyten Duyvil, and the legal center of the East River east of
Riker's Island. There's no particular value that I can see in the OSM
border in that area, which attempts to follow the coastline but is not
consistently conflated with it.

History of the boundaries in OSM is that they were imported from an old
TIGER (which was horrible - I've moved some boundaries Upstate by over a
km!)  and then pretty much just snapped to the shoreline by local mappers.
Neither appears to be correct, although the 'snap to the shoreline' option
wasn't as bad as TIGER!

The current OSM-modeled boundaries are also not quite consistent, for
instance the Queens-Bronx line is modeled with separate ways for county and
borough - https://www.openstreetmap.org/way/46936806 and
https://www.openstreetmap.org/way/113738569, and these are crossing ways.
(County and borough ARE supposed to be coterminous!)

I've already just about finished making all admin_level=6,7,8 boundaries
Upstate consistent with NYSGIS (and along the state lines, taking into
account the GIS of neighbouring states for the state boundaries. NY-CT was
pretty horrible before I started - I noticed that as it progressed, I was
seeing imported address points from both states lining up on the correct
sides of the border.

My experience with doing the rest of the state has been that when I look at
orthoimages, the TIGER boundaries weave about drunkenly while the NYSGIS
ones often align precisely with natural and cultural features - I can see
fence lines, the edges of parking lots, and the like, neatly underlaid with
respect to the NYSGIS version of the boundary.  TIGER also suffered from
runaway conflation, with boundaries snapping to the ends of nearby dead-end
streets, the alignment of driveways, pipelines, building walls and the
like.

 It's probably also worth noting that the Wiki
https://wiki.openstreetmap.org/wiki/Tag:boundary%3Dadministrative#Gluing
recommends AGAINST conflating political boundaries with natural features
except in the rare case where they are defined in such a way that if the
natural feature is altered, the boundary will move with it. That is not
generally true in New York: while gradual accretion and erosion may alter
cadastre, sudden avulsion or reliction will not. (Translation: If a river
moves or a hurricane alters the shoreline, and anyone cares about
the effect on land ownership, there will be lawsuits to determine it.)

If there is a serious objection, what I can do is preserve NYC's internal
boundaries but introduce little patch ways like
https://www.openstreetmap.org/way/1062389391 to preserve topology at the
margins. But I'd strongly prefer to purge the remaining vestiges of the
awful TIGER 2009 boundary import.
-- 
73 de ke9tv/2, Kevin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20220521/a3e4f12a/attachment.htm>