Colophon

Methodology & corrections

Passage is built on one rule: no number appears without a source, and when we get one wrong we say so here. This page is generated from the same data file the scenes read, so it can never quietly disagree with them.

The composite river

Passage follows an archetypal salmon on an archetypal river — a single continuous upstream journey stitched from several real basins. That composite is a storytelling device, and we label it rather than disguise it. Each basin switch is named on screen (the barrier landscape is a Pacific Northwest composite; the breakthrough is the Klamath; the decade of evidence is the Elwha), and the one live number is labeled exactly for what it is: adult Chinook counted at Bonneville Dam on the Columbia. We follow the upstream adult return, not the downstream smolt half of the lifecycle — a deliberate dramatic choice, with the ocean years addressed in the climate beat.

How we handle numbers

• Every figure is sourced. The build fails if any number a scene shows lacks a resolvable source and a verified date.
• Disagreement is shown, not resolved. Where two true framings exist (the Klamath 2025 run was both ~180% of the pre-season projection and ~61% of the long-term average), both appear.
• Estimates are flagged. A figure we could not yet trace to a primary source is marked with a ≈ est. and named below, pending a pinned citation before launch.
• Critiques are attributed and rebutted. The cost-and-doubt scene reports others' criticism as attributed opinion, always paired with the agency's own numbers — never as our factual verdict.
• Tribal data is consent-gated. This version uses public-domain agency data and already-published statements; tribal-collected data will not appear without consent.

Corrections applied

These are the errors an earlier draft carried, and what replaced them.

Figures still marked as estimates

These render with an ≈ est. mark until a primary citation is pinned. Listing them openly is the point — an honesty site should show its own open edges.

Full source ledger

Every confirmed figure Passage cites (40), with its source.

Changelog

• 2026-07-08Elwha sediment corrected: ~30M tons stored behind the dams, >20M tons released to the sea (the earlier "~10.5M metric tons released" was wrong). Both figures now stored and shown.
• 2026-07-08Dropped "largest dam removal in U.S. history" from the Elwha — the 2024 Klamath removal surpassed it. The Elwha is now framed by time (a decade of evidence), not scale.
• 2026-07-08Salmon SuperHwy shown as the org's 75% headline with the computed 76.1% (137/180) and a rounding note.
• 2026-07-08Cold Water Connection arithmetic made explicit: 32 opened + 43 remaining + 50 (WSDOT, separate) = 125 miles.
• 2026-07-08Washington culvert case disambiguated: the 2030 target is 90% of habitat; the extra up-to-$4B closes the 80→90% band, not "the last 10% of everything."
• 2026-07-08Economic value scoped to the PFMC ex-vessel figure; the $2.4B FEUS figure (which includes Alaska processed product) is explicitly not used.
• 2026-07-08Live counts wired to Columbia River DART (Bonneville adult Chinook), refreshed daily, with the USACE no-warranty disclaimer and the hatchery caveat attached to every appearance of the number.

Sources & licensing

Live counts: Columbia River DART (Columbia Basin Research, University of Washington), sourced from PSMFC and the U.S. Army Corps of Engineers — provisional, no warranty. The Elwha decade record draws on Pess et al. 2024 (Frontiers in Ecology and Evolution, CC BY 4.0) and public-domain USGS/NPS material. The Klamath reservoir-drain imagery is NASA public-domain Landsat. All charts are rebuilt from the underlying numbers; we never reproduce another organization's rendered chart or map.

Corrections and questions belong in the open. If you find an error, that is exactly the kind of thing this page exists to record.