[Imports] Status and progress: NYS DEC Lands reimport

Kevin Kenny kevin.b.kenny+osm at gmail.com
Tue Jul 19 15:55:41 UTC 2016


I just thought I'd make a quick follow-up here about the Long Island
pine barrens. I'm rather being a perfectionist.

A typical problem area can be seen at
http://www.openstreetmap.org/#map=17/40.89045/-72.62721 .  The
original imported data simply had an approximate property line, and
this line corresponds to the green-shaded region.

A later mapper tagged this region with 'natural=wood', and added
additional 'natural=wood' polygons coterminous with it.

The reimport saw the 'natural=wood' tagging as something that had to
be preserved, so kept the polygons while removing the tagging for
landuse=* and leisure=*. It then overlaid the imported area, tagged
'leisure=nature_reserve' and 'landuse=forest' as well as
'boundary=protected_area'.
http://www.openstreetmap.org/relation/6371616#map=17/40.89109/-72.62763
highlights the reimport.

A later mapping also glued a powerline to the preserve's southern
boundary. That took some manual repair.

This means that all along the border of the preserve, there are bits
tagged 'natural=wood' that are not in the preserve's boundary, and
other bits that are in the protected area but not tagged as woodland.
There's nothing fundamentally wrong with this - trees are no
respecters of property lines - but it looks untidy, particularly since
the original tagging was done based on the (incorrect) preserve
boundaries. (Much of the included area, too, is not woodland, but
scrub, or wetland, or brownfield.)

If I don't hear from a Long Island mapper, I think continuing the
process the way it began, with untidiness at the margins, is the best
that I'm going to be able to do. It's at least not undoing the work of
local mappers; with the exception that the protected areas are
redrawn, every bit of land is left tagged as I found it. It may or may
not be respecting their intent. Given that so much 'natural=wood' in
that area is tagging areas that aerial photographs make clear are NOT
woodland, it's hard to determine what was actually meant, beyond
'please paint this green on the map.'


On Mon, Jul 18, 2016 at 7:31 PM, Kevin Kenny
<kevin.b.kenny+osm at gmail.com> wrote:
> The reimport of NYS DEC Lands that I proposed this spring is now complete,
> except as noted below. Now would be a good time to check that I got it
> right.
>
> I am aware of the following issues that need to be resolved before I call
> the import 'complete.'
>
> (1) I have not imported a fair number of parcels in the Long Island pine
> barrens. The reason is that local mappers have in many cases reused the
> nodes marking the parcel boundaries as edges of polygons that specify
> further land uses, or land covers (wood, scrub, wetland, meadow, etc.) In
> many cases, these appear to have been attempts to tag for the renderer,
> since inspection of aerial photographs shows land cover inconsistent with
> the tagging (natural=wood on what is obviously a  brownfield site or a tidal
> flat, for example). This will require rather a lot of hand work to
> reconcile, and it's in an area where I don't have recent "boots on the
> ground" experience. If some local mappers from Suffolk County are willing to
> step forward and help me, that would be a tremendous relief. I can put up a
> .osm file with the DEC data as it should appear; other tagging (such as
> natural=*, landuse=*, related tags) obviously has to be reconciled with
> what's already there.
>
> (2) All the Long Island parcels, with the exception of the Ridge
> Conservation Area, need a mechanical edit to replace 'access=yes' with
> 'access=permit permit:website=http://www.dec.ny.gov/outdoor/7815.html.' I
> was unaware of this restriction when I started the import, and put Long
> Island on hold when I learnt about it, but there are a fair number of
> parcels already imported that need to be revised. I can come up with a
> script to identify them.
>
> (3) There are areas that comprise multiple separated parcels where a glitch
> in the import resulted in duplication of the multipolygon relation. Northern
> Montezuma Wildlife Management Area is the only one that I'm aware of, but
> I'm working up a script to audit the entire import and report duplicates.
>
> (4) Wildlife Management Areas exist that are not the current incarnation of
> the former State Game Farms. It would be a good idea to tag all Wildlife
> Management Areas with 'protection_object=habitat' rather than the current
> 'protection_object=hunting;fishing'. I fixed a couple of these manually, but
> want to do a mechanical edit to replace the rest.
>
> (5) I want to add leisure=nature_reserve to the landuse=aquaculture and
> boundary=protected_area that is present on the state fish hatcheries. This
> is not exactly a lie, since the land is protected, and if it is no longer
> used as a hatchery, it would revert to the default of 'state forest.' It's
> consciously tagging for the renderer, since there's otherwise nothing there
> that renders at all.
>
> (6) A number of parcels on the Hudson enjoy Federal as well as State
> protection, as the Hudson River National Estuarine Research Reserve. I need
> to finish researching which those are, and create a super-relation for them.
>
> (7) Many website=* links are dead or go to incorrect places. I have a script
> written already to enumerate the parcels that link to a website whose title
> differs significantly from the parcel name. These need to be inspected and
> better links found, if possible. (Failing anything else, links to the
> correct DEC regional office.) This is work that I could share out fairly
> easily, if anyone else wants to participate.
>
> (8) I have wikipedia= links for only a few of the most prominent sites. I'm
> sure that there are others with a Wikipedia presence, and searching for them
> could be partially automated. This is a low-urgency job, and again could be
> shared out pretty easily.
>
> On the whole, this reimport went considerably better than I expected. There
> were a few glitches that had to be corrected by stopping the import, doing
> mechanical edits, and resuming the import from where it left off, but on the
> whole the process was extremely smooth. In the vast majority of cases, there
> was no incompatible editing done since the initial import, so it was merely,
> "out with the old, in with the new." A check in JOSM for conflicting data
> took no more than a few seconds. (The other end of the spectrum was that a
> great many islands in the Adirondacks had to be drawn and added to their
> respective lakes. This got quite tedious indeed.)
>
> Thanks to all who helped with this!
>
> Kevin



More information about the Imports mailing list