Methodology and data sources

Trail data comes from the Forest Service TrailNFS_Publish layer on the agency's Enterprise Data Warehouse map server. The layer holds 86,410 trail segments. It was last read on 2026-09-17. Data refreshed September 17, 2026.

From segments to trails

The layer records segments, not trails. A segment carries a trail name, a beginning and ending milepost and the administrative unit it sits in. This site groups segments by trail name and by the forest in the first four digits of the administrative organization code, so a trail that crosses from one forest into another appears once per forest. Segments with no name, a placeholder such as UNNAMED, or a bare number or code in place of a name are left out. Segments with no geometry are left out. The result is 45,962 named trails.

Length is the sum of the segment lengths the forest recorded, which the data dictionary defines in miles as the ending milepost minus the beginning milepost. Where a segment length is blank the GIS-measured length is used instead. Trail class, surface, grade, tread width and accessibility are the set of values across the segments. A use is allowed when any segment allows it. The Forest Service records a use as managed, accepted, discouraged or restricted. The first three mean the use is legal on the trail and are counted as allowed. Restricted is not.

The start and end of a trail are the two ends of the longest connected chain of its segments. When the segments do not join into one chain, they are the two ends that lie farthest apart. Which end a person usually starts from is not in the record. States come from testing the line's vertices against Census state boundaries.

Which trails get a page

The host allows 20,000 files. Every trail with a national designation, every trail evaluated as accessible and every trail of at least 4.5 miles has its own page, which is 10,720 pages. The other 35,242 named trails appear as rows on their forest and state pages and in the open-data CSV, with the same facts.

Other sources

• Forest names come from the Forest Service Administrative Forest Boundaries layer. Forest home pages come from the recreation site URLs the Forest Service publishes for each forest.
• Trailheads and Forest Service campgrounds come from the Forest Service Recreation Opportunities layer. A trailhead is matched when it lies within a third of a mile of the trail line. Campgrounds are listed within five miles.
• Wilderness boundaries come from the National Wilderness Preservation System layer published by Wilderness Connect at the University of Montana. A trail is in a wilderness when its line crosses the polygon. The share of the line inside is reported.
• Campgrounds with pages on Campsite Atlas, park units on Park Atlas and National Register places on Register Atlas come from those sites' own indexes. Park units are listed within 25 miles and register places within 10.
• Wildfire incidents come from the interagency WFIGS feed and are matched to trail midpoints at build time.
• Weather, alerts, air quality, river flow and earthquakes in the sidebar are fetched by the browser from the National Weather Service, EPA AirNow and the US Geological Survey at the trail midpoint.

Known limitations

• Forests publish the layer at three levels of detail. About one segment in sixteen is a centerline only, with no class, surface or use recorded. Those trails show what the forest published and nothing more.
• Grouping by name merges trails that share a name inside one forest, and splits a trail whose name is spelled differently from one segment to the next.
• Lines are simplified to about 100 metres for the maps and the geometry tests. Short trails can lose their shape.
• Elevation gain is not computed. It is planned for a later build.
• Conditions, closures and hazards are not tracked. The forest's official page is linked from every trail page for that.