[Talk-us] National Forest boundaries

[stevea]

Adam Franco <adamfranco at gmail.com> writes (about my 1, 2, 3 post potentially defining NF MPs, now clarified that 1 isn't "all enclosing")
> I think this is correct:.


He continues:
> If there is consensus on dropping (3), then a system for mapping NFs as
> (1-2) should be possible to figure out. That said, how that OSM object is
> assembled and tagged may be tricky. In the Green Mountain National forest
> the (1-2) area contains a large mix of areas with different protections...
> Some of these child boundaries would have their own names and additional
> tags, others not.

Exactly!  I'm not yet ready to say 3) should be dropped, though I strongly lean in that direction, as I think it's unnecessary / superfluous given how we map "actual" data, not necessarily "Congressional" data, whatever that means.  Let's allow this list to concur and / or wider consensus to emerge about whether 3) can be clearly articulated enough as "here's how we should implement these data in the context of a well-crafted NF multipolygon," or whether it's not logically / geometrically necessary for OSM to denote and should be "dropped."  We'll eventually get there; we do inch closer.  Especially as it seems to be emerging that OSM can (and does) well-represent NFs with completely OSM-conforming multipolygons of the sort that I describe with 1) and 2), even while I/we look for additional guidance on 3).  Here's something I'll throw against the wall and see if it sticks:  maybe 3) (Congressionally-defined boundary) is a sort of crutch, a "nice to have," but not strictly logically / geometrically necessary except as a rough outline sketch of this NF (for Congress-critters, for low-zoom maps).  It seems we're there, but again, I solicit clarity on 3) here and now.

> I would imagine that the parent NF object that has the name
> "Green Mountain National Forest" would contain members that had
> protect_class=6 (resource extraction), protect_class=1b (wilderness),
> protect_class=5 (recreation areas, Appalation Trail corridor), etc.

Again, yes.  To clarify, the NF object itself (the multipolygon's tags, I'd discourage calling this "parent" as it means something else in the context of relations and super-relations and we shouldn't confuse those) would have the name=Green Mountain National Forest + protect_class=6 tags (plus others, like operator=USFS).  AND the additional members "associated with the NF" (like wilderness areas which are "within the NF") would be separate polygons with role inner as members of this NF relation, but ALSO with their OWN tags (like protect_class=1b and name=Breadloaf Wilderness).  Yes, this makes "holes" of wilderness inside of the NF, but think about it:  if the "whole thing less inholdings and stuff that's different" (outer minus inners) really deserves protect_class=6 AND the "inners that are wilderness" are tagged with protect_class=1b, well, we've got it!  Sure, doing it like that makes it logically appear (and maybe actually be) that wildernesses are excluded from the NF, but in the sense by which we tag them, they ARE excluded, even as they are surrounded by something with a "lesser" protect_class.  Plus, it is the same agency tagged both on the multipolygon (for the outer) and its member inners.  They are logically excluded by being inners, but because the MP is tagged operator=USFS (and so should be the inner wildernesses), we "add them back in," at least for their operator, by virtue of that tag being on the inners (But being differently tagged with protect_class=1b, as they should be).  Whew!

> I'm not sure what tagging would be appropriate for the NF object itself
> maybe these as a starting point?
>    - name=*
>    - boundary=national_park
>    - operator=US Forest Service

That IS a good starting point, for an exposition I recommend our wiki on this topic:
https://wiki.openstreetmap.org/wiki/United_States/Public_lands#Agriculture_Department_.28USDA.29_National_Forests_.28USFS.29.2C_National_Grasslands.2C_Special_Biological_Areas
(Full disclosure, I'm a significant author of this wiki, even as I and other authors earnestly seek wider contributions to it).  There, we say:

	• boundary=protected_area
	• protect_class=6
	• protection_title=National Forest
	• ownership=national
	• operator=United States Department of Agriculture, Forest Service

Regarding the subtopic (in this context) of "Ranger Districts," I think that can be accommodated with polygons of the particular areas that make up the MP of the NF and naming them accordingly.  It might take some work on the part of an intrepid OSM mapper to do this, as I'm not sure the way the USFS publishes the geo data of the NFs these are quite delineated "by Ranger District," but it could be done.  And maybe it should be, I think it would be a nice thing to map.

Hey, it's a TALK page.  We're TALKING.  It sometimes takes quite a few words to do that.

SteveA