Open Data & Attribution

Last updated: August 5, 2026.

OpenStreetMap trail data

TrailAware uses map data from OpenStreetMap to fill gaps in trail coverage, alongside trail data we hold in our own right. Where a trail comes from OpenStreetMap, TrailAware’s collection of that trail’s geometry is a database derived from OpenStreetMap and is made available under the Open Database License (ODbL) 1.0 — the same licence OpenStreetMap’s own data carries. Map data © OpenStreetMap contributors.

How we build this collection (ODbL section 4.6(b))

The Open Database License requires us to either publish the derived database itself, or describe the method used to alter OpenStreetMap’s data, free of charge, over the internet. This section is that description. It covers OpenStreetMap-derived trail geometry only — it does not extend to TrailAware’s community data, favorites, condition reports, AR calibration data, or partner-venue content, none of which is built from OpenStreetMap.

• Scope. We pull data park by park, for U.S. National Park Service units carrying the official “National Park” designation. For each park we resolve its official boundary in OpenStreetMap — a relation tagged boundary=national_park or boundary=protected_area whose name matches the park’s full name — and query only inside that boundary.
• Feature types. Within the boundary we extract ways tagged highway=path, footway, track, bridleway, steps, or cycleway. We also pull a small set of facility points near trails — parking, toilets, drinking water, shelters, ranger stations, viewpoints, information points, huts, picnic sites, campsites, and trailheads — tagged with the matching amenity, tourism, or highway=trailhead values.
• Named trails. Of the matching ways, only those carrying an OpenStreetMap name tag become a trail record you can see and rate; unnamed connector ways are read as part of the same query but are not currently added to the trail collection.
• Stitching. OpenStreetMap usually represents one physical trail as several short way segments that meet end to end. We join segments that share an endpoint into a single continuous line, then group same-named lines that lie within roughly 800 metres of each other into one named trail. Same-named lines further apart are kept as separate trails, since generic trail names (like “Loop Trail”) repeat across a park. Joined segments shorter than 150 metres are treated as mapping stubs and dropped.
• Simplification. The resulting line for each trail is simplified with the Ramer–Douglas–Peucker algorithm and capped at 500 points, so the geometry stays light enough to render on a map and to request elevation for. This changes point density, not the trail’s shape or location.
• Tag normalization. OpenStreetMap’s surface, sac_scale, and mtb:scale tags are mapped onto TrailAware’s own difficulty (Easy / Intermediate / Advanced / Expert) and surface vocabulary — the same vocabulary already used for trails we hold outside OpenStreetMap, so trails read consistently regardless of source.
• Elevation. OpenStreetMap does not carry elevation for these ways. Where TrailAware shows an elevation profile for one of these trails, that value is added afterward from a separate, non-OpenStreetMap elevation source (see our Privacy Policy for our data providers).
• Matching against our own data. Each candidate OpenStreetMap trail is compared against nearby trails TrailAware already holds, within a roughly 2 km area. It counts as a match when the name matches and at least 30% of its points fall within 25 metres of the existing trail’s line, when at least 70% of its points do so regardless of name, or when the name matches and the two trails start within 250 metres of each other. A match keeps the existing trail as the one users see and rate; the OpenStreetMap record is kept and tagged as a duplicate rather than deleted, so the OpenStreetMap-derived portion of the database stays intact and identifiable rather than merged away.

The Overpass query

We fetch OpenStreetMap data from the public Overpass API, one park at a time, identifying our requests with a descriptive user agent and contact address as OpenStreetMap’s API usage policy asks. First we resolve the park’s boundary:

[out:json][timeout:180];
rel["boundary"~"^(national_park|protected_area)$"]["name"="<park full name, e.g. Yosemite National Park>"];
out ids tags;

Then, using the OpenStreetMap area id computed from that boundary’s relation id, we fetch the trail geometry:

[out:json][timeout:300];
way(area:<area id from the boundary query>)["highway"~"^(path|footway|track|bridleway|steps|cycleway)$"];
out tags geom;

And the nearby facility points:

[out:json][timeout:300];
(
  node(area:<area id>)["amenity"~"^(parking|toilets|drinking_water|shelter|ranger_station)$"];
  node(area:<area id>)["tourism"~"^(viewpoint|information|alpine_hut|wilderness_hut|picnic_site|camp_site)$"];
  node(area:<area id>)["highway"="trailhead"];
  way(area:<area id>)["amenity"~"^(parking|toilets)$"];
);
out center tags;

Get the altered data

Instead of this description, you may request the current OpenStreetMap-derived trail database as we hold it — free of charge, over the internet — by emailing [email protected] or using the licensing & business form on our Contact page. We will send an export or a link to one.

National Park Service content and photos

Some park descriptions, alerts, and photos come from the National Park Service (nps.gov). Text and photos authored by NPS staff are U.S. Government works and are in the public domain. NPS also hosts photos on nps.gov that were contributed by outside photographers who keep their own copyright; we exclude any photo whose credit line names someone other than NPS, rather than assume it is free to use. Where we republish NPS public-domain material, no claim is made to original U.S. Government works. See the National Park Service’s own disclaimer and usage policy.

Wikipedia extracts

Some place descriptions are a short extract from the lead section of the relevant Wikipedia article. Wikipedia text is licensed CC BY-SA 4.0. Wherever we use it we credit “From Wikipedia, CC BY-SA 4.0” with a link to the specific article.

Wikimedia Commons images

Some photos come from Wikimedia Commons. We only use files that carry a machine-readable free licence — public domain, CC0, CC BY, or CC BY-SA; files marked non-commercial, no-derivatives, or carrying an additional usage restriction are excluded automatically. Each image we use is credited individually, next to the image, with the creator’s name, the licence name, and a link to the licence and to the file’s page on Commons.

How we credit all of this

Wherever content from one of these sources appears, we credit it the same way: the source, the licence it carries, and a link to the original — next to the content itself for anything credited to an individual creator (Wikimedia Commons photos), and in the site footer or an about/credits screen for attribution that covers a whole collection (OpenStreetMap map data, NPS public-domain material). That matches the placement OpenStreetMap’s own guidance treats as sufficient: a visible, linked credit that does not need to repeat on every single view.

Questions or data requests

For anything on this page — a request for the current OpenStreetMap-derived dataset, a correction, or a licensing question — email [email protected].