Re: [Talk-us] US Bureau of Land Management Boundaries

[Martijn van Exel]

Brad — 

My reference to Gaia GPS was meant to illustrate that third party apps are 
perfectly capable of overlaying data from various sources. Just because a data 
source exists doesn’t mean that it should be in OSM. On the ground 
verifiability has always been the gold standard for OSM, and I feel strongly 
about keeping it that way. We have departed from that gold standard somewhat 
already, especially when it comes to administrative boundaries.

Another argument for keeping this data out of OSM comes to mind. In many 
places, OSM has come to be the most complete map out there. OSM data is now 
used by major companies and organizations, and is therefore increasingly seen 
as ’the truth’. With that comes scrutiny and responsibility. We have already 
seen the project drawn into political disputes that are not ours to take a 
stance in. To give one recent example, our representation of the Ukraine / 
Crimea border recently drew sharp criticism and unneeded negative attention to 
OSM that could easily have been avoided. I’m not saying that public lands 
boundaries in the U.S. are to be compared with international conflicts, but 
public lands are in fact under heavy scrutiny here in the West, and 
incorporating these boundaries into OSM would just generate another attack 
vector we can do without.

If you want a Garmin map that incorporates both OSM data and boundary data from 
BLM or other federal or state sources, there are great ways to accomplish that 
goal, and I would be happy to help with it.

Martijn

> On Jan 6, 2019, at 8:50 AM, brad  wrote:
> 
> Martijn,  Gaia is not available on a Garmin, or on a PC.  It also costs $40 a 
> yr.   Why do you trust Gaia as an authoritative source?   How often do they 
> update from government sources?   BLM boundaries do not change very often.  
> Probably less often than city/town boundaries. For an authoritative 
> source, I have national forest maps that are 10 - 20 years old.  A download 
> today from a federal database is way better than that and in 5 years will 
> probably still be just as good.In relatively sparsely populated areas, on 
> the ground verification does not work as well as it does in the city.If 
> we make OSM more useful for more people then more folks will get involved.
>   

___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US Bureau of Land Management Boundaries

[OSM Volunteer stevea]

On January 6, 2019 at 7:50:44 AM PST, brad  wrote:
> 
> Joseph,   I'm not stuck on class 27, but as you say, that fits the definition 
> on the wiki.   I should probably look for other specific protection in the 
> attributes and translate that somehow.   Mostly it's just grazing and 
> recreation land.   Anything such as wilderness or monument would definitely 
> be tagged as such.

I agree with this approach, especially "look for other...attributes and 
translate (them)."  However, this is not something Brad "should probably" do, 
it MUST be done to correctly import these data:  each parcel must be examined 
as to its landuse (in the generic sense, not the OSM tag) and assigned an 
appropriate protect_class, especially if it is not 27.  The protect_class key 
may not render today (though, Carto will hopefully "fix" this with likely 
progressively improving methods in the future), that is a separate issue I 
specifically point out so additional tags (such as leisure=nature_reserve) do 
not get added superfluously ("tagging for the renderer") to make them "appear." 
 Beware this slippery slope, knowing that even a perfect, completed BLM import 
will (today) be essentially invisible in virtually all renderers.  The data 
being in the map is a good goal, even while rendering them can be put off until 
another day (though, not forever).

> Martijn,  Gaia is not available on a Garmin, or on a PC.  It also costs $40 a 
> yr.   Why do you trust Gaia as an authoritative source?   How often do they 
> update from government sources?   BLM boundaries do not change very often.  
> Probably less often than city/town boundaries. For an authoritative 
> source, I have national forest maps that are 10 - 20 years old.  A download 
> today from a federal database is way better than that and in 5 years will 
> probably still be just as good.In relatively sparsely populated areas, on 
> the ground verification does not work as well as it does in the city.If 
> we make OSM more useful for more people then more folks will get involved.

As a segue from my recent comments on USA rail being about 40% done (over a 
decade since their nationwide TIGER import), with such challenges (importing 
nationwide data such as BLM boundaries) come great responsibilities.  To 
repeat:  we imported "all" (that TIGER had) of US railroad data and here we 
are, eleven, twelve years later at about 40% completion of reviewing, improving 
and reporting on their status.  Such nationwide tasks (in the USA) are 
Herculean efforts, though breaking things up into wikis / efforts at a state 
level has proven effective (if relatively slow, it does make logical sense 
given state DOTs create rail inventory / planning reports every so often, which 
help a lot).  Should this BLM data import progress, Brad needs to know how 
large an elephant this is to eat.  I began similar importation of national 
forest (and wilderness, national grassland...) data in California between 
2012-3 but abandoned doing so, as the effort simply overwhelmed my ability to 
either do this myself or do it with the coordinated effort of other OSM 
volunteers.  I cannot emphasize this enough:  to do and manage these kinds of 
national-level data management tasks is an absolutely huge undertaking and I 
speak from extensive experience at either attempting or (partially, 
successfully, unsuccessfully) completing two or three of them (national rail, 
national bicycle routes, NF/Wilderness/BLM/other federal lands).

> Michael,   You bring up some good questions which I don't have the answer for 
> yet.   I would get started with what you call the low road, state sized or 
> smaller pieces at a time.  A quick look at the boundaries around me show none 
> that follow a watercourse or a ridge, they are all straight lines and and 
> square corners.  The extraneous ways at state boundaries look like artifacts 
> from cutting up a larger database into state size chunks.  There was no 
> polygon, or a skinny polygon associated with those artifacts.  I'm guessing 
> that there is BLM land in the adjacent state.

I enthusiastically encourage an initial pilot project of a single 
state-at-a-time's worth of data.  It is far easier to scale up (or abandon) 
something you can bite and chew (and swallow and digest) rather than try to 
scale down a disastrous import that is so large you (and OSM) choke on it.

> Dave, Thanks for being a voice of reason!  

I also agree with Dave's and Brad's assertions that these data belong in OSM.  
Publicly-owned BLM lands afford numerous recreational, educational and other 
opportunities, similar to leisure=park, leisure=nature_reserve and related 
areas.  Denoting where these are with recent federally published data is in 
complete harmony with other sorts of boundaries in OSM.  But there is wishing 
or agreeing that the data belong, then there is doing a high quality job of 
importing and maintaining them.  The former is relatively easy, the latter is 

Re: [Talk-us] US Bureau of Land Management Boundaries

[Mark Wagner]

On Sat, 5 Jan 2019 21:19:10 -0600
Ian Dees  wrote:

> Hi Brad, thanks for proposing this import and posting it here.
> 
> I would strongly prefer that we not import boundaries like this into
> OSM. Boundaries of all sorts are almost impossible to verify with
> OSM's "on the ground" rule, but BLM boundaries in particular are such
> an edge case (they have no other analog in the world, really) and
> almost never have apparent markings on the ground to check. Since
> these boundaries aren't visible, this data can never be improved by
> an OpenStreetMap contributor. The boundaries are defined by the
> government, and any sort of change to them would make them diverge
> from the official source.

I don't know about BLM land in general, but I was able to verify a
large portion of Fishtrap Recreation Area
(https://www.openstreetmap.org/relation/6154436) by tracing fence lines
and looking for "No Trespassing" signs.  I suspect most of the other
BLM rangeland can be mapped this way as well.

-- 
Mark

___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us

Re: [Talk-us] US Bureau of Land Management Boundaries

[brad]

Joseph,   I'm not stuck on class 27, but as you say, that fits the 
definition on the wiki.   I should probably look for other specific 
protection in the attributes and translate that somehow.   Mostly it's 
just grazing and recreation land.   Anything such as wilderness or 
monument would definitely be tagged as such.


Martijn,  Gaia is not available on a Garmin, or on a PC.  It also costs 
$40 a yr.   Why do you trust Gaia as an authoritative source?   How 
often do they update from government sources?   BLM boundaries do not 
change very often.  Probably less often than city/town boundaries. 
For an authoritative source, I have national forest maps that are 10 - 
20 years old.  A download today from a federal database is way better 
than that and in 5 years will probably still be just as good.    In 
relatively sparsely populated areas, on the ground verification does not 
work as well as it does in the city. If we make OSM more useful for more 
people then more folks will get involved.


Michael,   You bring up some good questions which I don't have the 
answer for yet.   I would get started with what you call the low road, 
state sized or smaller pieces at a time.  A quick look at the boundaries 
around me show none that follow a watercourse or a ridge, they are all 
straight lines and and square corners.  The extraneous ways at state 
boundaries look like artifacts from cutting up a larger database into 
state size chunks.  There was no polygon, or a skinny polygon associated 
with those artifacts.  I'm guessing that there is BLM land in the 
adjacent state.


Dave, Thanks for being a voice of reason!

Brad


On 1/6/19 3:36 AM, Dave Swarthout wrote:

Ian Dees wrote:

>"Those things shouldn't be in OSM either"

Are you implying that because such boundaries (National Forests, 
National Parks and National Wildlife Refuges) are non-verifiable by 
OSM mappers they don't belong in OSM? If so, wow!


I live in Alaska where about 60% of the land area is included in one 
or the other of these categories and I firmly believe they must be 
included in OSM. No, the boundaries aren't verifiable nor are they 
particularly accurate but actually nothing we put into OSM is accurate 
within several meters at best. If I add a National Wildlife Refuge 
having a total area of thousands of sq. kilometers and boundary errors 
amount to a few dozen sq km, for example, that doesn't bother me one 
bit. At least it's there for people to see. I'm really not looking for 
super accuracy; what I want are the visible outlines of those 
protected areas, be they rough estimates or not. OSM makes no warranty 
concerning accuracy. If you want to build a home near a NP or NWR 
boundary you'll need a surveyor; our rough boundaries won't serve for 
that purpose.


Can you elaborate on your statement, please?

Dave

On Sun, Jan 6, 2019 at 2:05 PM Martijn van Exel > wrote:


Brad — I make use of BLM / NPS / NF boundary data a lot too. I use
Gaia GPS for this, which overlays this data nicely with what’s in
OSM[1]. There are lots of other outdoor apps that do the same. I
prefer this data live outside of OSM as well for similar reasons
as Ian stated. Knowing whether land is public or private or
whether it’s inside or outside a NP, is important to me when I’m
in the outdoors. I would much rather rely on an authoritative
definition of these boundaries, than on whatever happens to be in
OSM. Since there is no on-the-ground verifiability, boundary data
is prone to growing stale, as you can see happening with census
place boundaries. Unreliable data in this case is worse than no
data at all.
If you’re looking to make a great impact on the map as an outdoors
user, I would suggest mapping things you know and things you
observe when you’re out there. Countless times have I been out in
the middle of nowhere, to find that some mapper before me added a
landmark, a water source, or something else that really helped me.
That is what I like to pay forward.

Martijn

[1] https://www.gaiagps.com/offroad/#maps


On Jan 5, 2019, at 8:43 PM, brad mailto:bradha...@fastmail.com>> wrote:

Ian,
I want to import this data because I think its important for a
complete map.   We have national forest, wilderness  and national
park boundaries in OSM!   This is no different. If you look at
many maps they show all of them.

I'd like it to show up on any map that I use.   I'm working on a
'better' version for garmin using mkgmap.   I hope it gets
rendered with OpenAndroMaps too.   I haven't used the onine
osm.org  map very much.

I am excited to participate and improve OSM and in my opinion
this is a big gap in the OSM database.   Where I live, we don't
use OSM for building footprints, we use it to find our way in the
national forest, the BLM land and the national parks.   It's very
useful t

Re: [Talk-us] US Bureau of Land Management Boundaries

[Dave Swarthout]

Ian Dees wrote:

>"Those things shouldn't be in OSM either"

Are you implying that because such boundaries (National Forests, National
Parks and National Wildlife Refuges) are non-verifiable by OSM mappers they
don't belong in OSM? If so, wow!

I live in Alaska where about 60% of the land area is included in one or the
other of these categories and I firmly believe they must be included in
OSM. No, the boundaries aren't verifiable nor are they particularly
accurate but actually nothing we put into OSM is accurate within several
meters at best. If I add a National Wildlife Refuge having a total area of
thousands of sq. kilometers and boundary errors amount to a few dozen sq
km, for example, that doesn't bother me one bit. At least it's there for
people to see. I'm really not looking for super accuracy; what I want are
the visible outlines of those protected areas, be they rough estimates or
not. OSM makes no warranty concerning accuracy. If you want to build a home
near a NP or NWR boundary you'll need a surveyor; our rough boundaries
won't serve for that purpose.

Can you elaborate on your statement, please?

Dave

On Sun, Jan 6, 2019 at 2:05 PM Martijn van Exel  wrote:

> Brad — I make use of BLM / NPS / NF boundary data a lot too. I use Gaia
> GPS for this, which overlays this data nicely with what’s in OSM[1]. There
> are lots of other outdoor apps that do the same. I prefer this data live
> outside of OSM as well for similar reasons as Ian stated. Knowing whether
> land is public or private or whether it’s inside or outside a NP, is
> important to me when I’m in the outdoors. I would much rather rely on an
> authoritative definition of these boundaries, than on whatever happens to
> be in OSM. Since there is no on-the-ground verifiability, boundary data is
> prone to growing stale, as you can see happening with census place
> boundaries. Unreliable data in this case is worse than no data at all.
> If you’re looking to make a great impact on the map as an outdoors user, I
> would suggest mapping things you know and things you observe when you’re
> out there. Countless times have I been out in the middle of nowhere, to
> find that some mapper before me added a landmark, a water source, or
> something else that really helped me. That is what I like to pay forward.
>
> Martijn
>
> [1] https://www.gaiagps.com/offroad/#maps
>
> On Jan 5, 2019, at 8:43 PM, brad  wrote:
>
> Ian,
> I want to import this data because I think its important for a complete
> map.   We have national forest, wilderness  and national park boundaries in
> OSM!   This is no different.   If you look at many maps they show all of
> them.
>
> I'd like it to show up on any map that I use.   I'm working on a 'better'
> version for garmin using mkgmap.   I hope it gets rendered with
> OpenAndroMaps too.   I haven't used the onine osm.org map very much.
>
> I am excited to participate and improve OSM and in my opinion this is a
> big gap in the OSM database.   Where I live, we don't use OSM for building
> footprints, we use it to find our way in the national forest, the BLM land
> and the national parks.   It's very useful to know what is public or
> private land.
>
> Brad
>
> On 1/5/19 8:19 PM, Ian Dees wrote:
>
> Hi Brad, thanks for proposing this import and posting it here.
>
> I would strongly prefer that we not import boundaries like this into OSM.
> Boundaries of all sorts are almost impossible to verify with OSM's "on the
> ground" rule, but BLM boundaries in particular are such an edge case (they
> have no other analog in the world, really) and almost never have apparent
> markings on the ground to check. Since these boundaries aren't visible,
> this data can never be improved by an OpenStreetMap contributor. The
> boundaries are defined by the government, and any sort of change to them
> would make them diverge from the official source.
>
> But having said that, I'm curious why you wanted to import this data? Did
> you want to have it show up on the osm.org map? Are you trying to build a
> custom map? Or are you excited to participate and improve OSM? If it's the
> latter, there's lots of other data that is a better fit to import into OSM:
> address points and building footprints come to mind, for example.
>
> -Ian
>
> On Sat, Jan 5, 2019 at 9:03 PM brad  wrote:
>
>> I'd like to import BLM (US Bureau of Land Management) boundaries into
>> OSM.This is not an automated import as you can see from my workflow.
>>
>> Workflow:
>> Download shape file from PADUS (1 state at a time):
>> https://gapanalysis.usgs.gov/padus/data/download/
>> Load into Qgis and filter for BLM boundaries
>> Clean up as necessary (there are some extraneous ways at state
>> boundaries & elsewhere)
>>
>> Convert to OSM with ogr2osm and the following tags
>>  tags.update({'type':'boundary'})
>>  tags.update({'boundary':'protected_area'})
>>  tags.update({'operator':'BLM'})
>>  tags.update({'ownership':'national'})
>>  tags.update({'protect_clas

Re: [Talk-us] US Bureau of Land Management Boundaries

[Michael Patrick]

> Clean up as necessary (there are some extraneous ways at state
> boundaries & elsewhere)

BLM manages about 10% of the total area of the United States, and those
areas historically have had the least resources dedicated in terms of
mapping. Also, the BLM data is an amalgamation of data from corporate (
mining usually ), various government agencies at local, county, state, and
federal levels. This data from other sources may have been incorporated
piece meal over many decades, and not updated since ( like 'way back' US
Census Tiger line files ). Some of these conflation issues are discussed in
this article: https://www.esri.com/news/arcuser/0110/accuracy-precision.html

With any of these datasets, one should read ( and understand ) the metadata
associated with the file. For example:
https://pubs.usgs.gov/ds/821/downloads/metadata.txt - this had a fairly
rigorous QA to fairly current imagery. So it might actually be much better
than the existing data in OSM that came from the TIGER import ( which was
in turn derived by the US Census from much older data ). The opposite
situation can also occur - some BLM data was manually from paper maps.

Exactly what is the decision tree when these issues are encountered? A read
through that metadata can give an idea of how they parsed out the issues,
for that theme.

For instance, frequently area BLM boundaries follow watercourse, and I do a
check against the best available orthophotography and the USGS  3DEP (
https://www.usgs.gov/news/new-elevation-map-service-available-usgs-3d-elevation-program
). Similarly with ridge lines. Boundaries with USDA National Forest and
state forest frequently interweave. Also, one theme ( layer ) for an area
might be excellent and the other just crap - BLM manages for lots of
different uses, and some just get more attention than others.

There's a high, a middle, and a low road here - most likely for low end
expediency you'd just leave any existing OSM lines in place even if they
varied considerably, and just import any BLM information if it didn't
already appear in OSM, in the middle you'd pick the best and adjust the OSM
data where it was considerably different, and the high would be a thorough
effort.

At one point there was a plug-in for openJUMP called RoadMatcher that might
be useful, last I looked the tools in qGIS were lagging. ESRI sells a
'personal edition' version of ArcGIS for a $100 that has extensive tools
for these tasks, it becomes worth it in about two hours.

Just curious ... what makes those 'extraneous ways at state boundaries',
extraneous?

> Clean up as necessary...

So, I guess I'd be interested in how you get to the necessary. The BLM data
covers an area about half the size of Europe.

Michael Patrick
Data Ferret
___
Talk-us mailing list
Talk-us@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-us