Methodology

Every guide on this site is built from a structured manifest, validated against a schema, fact-checked against named sources, and rendered from the same template. This page documents that process so you can judge the work.

Per-fall data sources

Each guide's manifest pulls from a fixed hierarchy of authoritative sources:

• Land manager pages for hours, fees, parking, swim policy, dog rules, and access alerts. USFS recreation listings, NPS park pages, state-park pages, BLM recreation pages, and city/county park pages.
• USGS National Water Information System for live discharge (cubic feet per second), the 30-year historical median, 75th and 90th percentiles, and the maximum recorded reading. 14 of 52 guides currently have a paired discharge gauge wired to the page.
• NOAA / National Weather Service gridpoint forecast API for the build-time three-day forecast strip on each guide.
• State geological surveys and the USGS Mineral Resources Online system for bedrock identification and formation history.
• Wikipedia and Wikidata for entity disambiguation, etymology, and historical references, cross-checked against primary sources.
• AllTrails for crowd-sourced trail conditions and recent visitor reports (used as a recency check, not a primary source).

Photos

Visuals on each guide are editorial composites grounded in verified reference photography of the specific waterfall. The approach gives every guide the same standard of visual coverage and supports falls that are remote, seasonal, or under restricted access where a single documentary photograph would be misleading. Reference images are licensed (primarily Wikimedia Commons with usable license metadata, occasionally land-manager media), and every image used as a source is recorded with its license and creator. Composites are reviewed for visual identity and factual consistency before publishing. Factual claims on the page never come from the images themselves; they come from the sourced research above. The editorial framing for visuals is documented on the editorial standards page.

Build process

The build pipeline (Python plus Jinja2) validates every manifest against a schema, fetches live USGS and NWS data at build time, computes haversine-sorted related-falls links between guides, and renders the page from a single template. The build is idempotent and rebuilds the full site in about 90 seconds.

Each guide ships with JSON-LD structured data for Article, TouristAttraction, BreadcrumbList, and FAQPage. The FAQPage block indexes the eight per-fall cluster questions for AI answer engines and Google rich-result eligibility.

What we deliberately don't do

• No fake precision. If a height is contested between sources, we say so on the page rather than picking one.
• No invented USGS readings. A guide only shows a live discharge chip if a real paired gauge exists in the manifest. Otherwise the page falls back to seasonal historical norms.
• No SEO-shaped padding. The long-form essay on each guide answers a distinct keyword cluster, not the same body text repeated with the fall name swapped.
• No affiliate links or sponsored placements. External links are source verification.

Contact

Methodology questions, accuracy disputes, or research collaboration: [email protected].