[Talk-us] US Hydrographic Data

Matthew Perry perrygeo at gmail.com
Thu May 1 20:20:53 BST 2008

On Thu, May 1, 2008 at 11:54 AM, Ian Dees <ian.dees at gmail.com> wrote:
> What do you think needs to be done to the NHD import script to make it
> "final" -- or at least in a state that we could start running imports for
> the whole country?
> I should be receiving the DVDs tomorrow or Saturday. Dave, what were your
> experiences importing the TIGER data? How should we proceed importing this
> data?

Good questions.  I think the header of the script outlines the
deficiencies pretty well so I'll just paste that verbatim:

Flowline:StreamRiver -> waterway:stream
Flowline:ArtificialPath (etc) -> waterway:canal
Flowline:Pipeline -> man_made:pipeline;type:water
Waterbody:SwampMarsh -> natural:marsh
Waterbody:Reservoir -> natural:water;man_made:reservoir
Waterbody:LakePond -> natural:water

- Uses GNIS_NAME as name tag if available.
- Applies oneway:true tag to streams and rivers to indicate flow direction

 - For polygon waterbodies, need to incorporate holes
 - incorporate all dbf fields
 - map the full NHD schema to osm features/tags
   (full desc of NHD schema:  http://nhd.usgs.gov/chapter2/index.html )

The polygon holes (eg islands in lakes) are easy to identify in the
shapefiles but I don't know the OSM data model well enough to
incorporate them.

The import script currently only pulls out select features and only
maps select attributes to OSM tags. I feel like we should look over
the whole dataset and see what else might be of use. We should look at
it with an eye towards not only making maps but keeping the data
structure intact so that meaningful hydrologic modeling/routing can be
accomplished as well.

We would probably want to spot check a couple of hydrologically
diverse areas and make sure it handles everything properly (maybe take
a sampling of lake regions, swamps, urban areas, mountains, deserts,

Matthew T. Perry

More information about the Talk-us mailing list