# National Forest MVUM Highways

This is a data extract of National Park Service Trails from their
official sources. They are organized alphabetically by the park name
rather than by state, since some cross state lines, which messes up
the data extracts.

These files have been machine edited by the the 
[trails.py](https://github.com/hotosm/osm-merge/blob/main/osm_merge/utilities/mvum.py)
script, which is part of the [OSM
Merge](https://hotosm.github.io/osm-merge/) project. Documentation on
the conversion process is
[here](https://hotosm.github.io/osm-merge/highways/).

# File Types

For now all the files are in a single directory. Later I'll split all
these into tasks for the Tasking Manager.

* *.pbf
  * The OpenStreetMap data for this region
* *_NPS_Trails.geojson
  * Official National Park Service trail data
* *.geojson
  * The region boundary for ogr2ogr or osmium
* *.poly
  * The region boundary for osmconvert

# Bugs

This is not a complete list, as I have not gone through every national
forest data deeply enough. Just the ones in my area. As I stumble
across problems in the MVUM dataset, I'll add them to the list. If you
find some bad data, let me know and I'll add them to this list.

## Bad Reference Numbers

This of course makes conflation difficult, as the MVUM dataset is
considered the __official__ source for the reference numbers. It's
potentially possible to clean this up, by conflating with an
additional external dataset like topographical maps. These are easy to
find as the post conflation data has both, so it's really obvious. But
it slows down validation a little, and you have to conflate multiple
datasets.

### Manti-La_Sal National Forest

In the current MVUM dataset for this national forest, for some reason
a *5* or *7* has been prefixed to many of the IDs, making the reference
numbers wrong.

### Fishlake National Forest

In the current MVUM dataset for this national forest, for some reason
a *4* or *40* has been prefixed to some of the IDs, making the
reference numbers wrong.

### Dixie National Forest

In the current MVUM dataset for this national forest, for some reason
a *30* has been prefixed to all the IDs, making the reference numbers
wrong.

### Mount Hood

For some reason, some of the reference numbers have a *000* appended,
making the reference numbers wrong. This applies to paved roads, not
just remote jeep tracks.