[Talk-us] (Second attempt) Potential data source: Adirondack Park Freshwater Wetlands

[stevea]

Elliott Plack <elliott.plack at gmail.com> writes:
>I would argue that importing land-use that is difficult or tedious 
>to trace...encourage(s) local mapping...

I agree.  OSM user nmixter and I contributed to a comprehensive 
landuse import in Santa Cruz County California starting in 2009.  You 
can read the details at our County's wiki page, but over three 
versions and five+ years, this is now reasonably up-to-date with 
two-year-old (the latest) data.  While painstaking manual updating of 
over 3000 (multi)polygons was required, we (largely I) did it, 
because we should have.  Santa Cruz even won a Gold Star Award from 
BestOfOSM.org, one of only a handful of North American locations to 
receive this accolade.  From the site:  "...nearly perfect landuse!" 
And as newer landuse data become available, published by our County 
GIS Department, we (I, likely) will update/import these to version 4. 
That's what it takes, so that's what we do.

We have many active local contributors, including students from our 
local University of California campus via numerous (Computer Science, 
Environmental Studies...) classes, as well as the Transportation 
Department adopting OSM for the campus' official basemap at 
http://maps.ucsc.edu.  The map is plastic enough to accommodate all 
of these uses, AND be a useful academic tool at the same time.  Just 
have respect for the data, follow good, simple rules and have fun!

>1. It shows others that an area of the map has received some attention.

Yes, it certainly does!  However, a landuse import must be done well, 
with an eye towards quality, accuracy and even beauty, rather than 
simply be a bad excuse for "spilling large buckets of paint" 
(farmland, meadow, residential...) all over the map with 
poorly-chosen, giant landuse= (multi)polygons.  Please be careful not 
to do this; a landuse import is a balancing act.  For example, we 
know that our landuse import, which might result in what some think 
of a "zoning map" (it is much more than that) is really a first draft 
for much more comprehensive OSM data input in the future.  It 
ENCOURAGES these additional data entries by taking the map from 
"empty" to "rough (but accurate) sketch" to "now that others have 
gotten you started, please draw as much beautiful detail as you are 
able to."

>2. It produces "gaps," i.e. places where there are no wetlands or 
>water thus leaving a gray "hole" on the map, thus a mapper might 
>look there to add something, such as a camp site or some unknown 
>settlement.

In our County, this draws special attention to parks that haven't 
been named properly, so-called "special_use" parcels (some turn out 
to be landuse=industrial like a water tower or sewage treatment), 
sharpening up differences between residential farmland (so-called 
"live-on" family farms) vs. more commercial farmland like orchards, 
vineyards and row crops and it allows boundaries (e.g. of public 
lands) to be improved, TIGER roads to be called attention to with 
obvious errors so they can be improved, and more.

>3. It exposes inaccurate TIGER roads and tracks, as these typically 
>are on embankments through wetlands.

As I just said.  Sometimes TIGER roads and tracks follow a property 
boundary, and if a landuse import is accurate, it can be used to 
supercede the TIGER road or track, allowing obvious improvement.

>Importing landuse adds visual beauty to the map in places where it 
>would take hordes of volunteers to trace wetlands, and experts to 
>determine the wetland classification. I think this is a good idea.

And not just in wetlands, though I do agree.  In our world (2/3 of 
our County is wooded) it was found that a tag of natural=wood (where 
landuse=forest, or an active timberland) was appropriate for many 
parcels (such as special_use) that otherwise wouldn't render.  When 
superimposed with the simple task of tracing existing landuse=meadow 
polygons over these woods, forests and parks, the effect is both 
accurate and visually quite pleasing.  (See, for example, the rather 
pretty -- and accurate -- contrasts between wooded/forested areas and 
farms, parks and meadows at 
http://www.osm.org/#map=16/36.9712/-122.0778).

>I've fiddled around with NWI data and OSM and usually the result is 
>great. While I'd concede that landuses do change, so to do other 
>features that are imported often (buildings and such). A map is 
>always a snapshot in time, and for the most part land use within a 
>protected area is not subject to much in the way of change.

Likewise.  We (OSM volunteers) don't often talk about the importance 
of KEEPING UP the map after an import, but doing so is a seriously 
crucial component.  Thank you, Elliott, for calling this to our 
attention.  Some updates need relatively frequent updating, some 
almost none at all.

SteveA
California