Eclipse 2026 Prediction Registry

The prediction registry — sealed before the sky moves

On 2026-07-16, twenty-seven days before the eclipse, the substrate sealed a falsifiable prediction of the August 12 ionospheric response to V234 (ECLIPSE_2026_PREDICTION_PREREGISTERED) and to a fingerprint-pinned Form. Nothing about it can change without the fingerprint saying so.

The sealed claim, verbatim

Sealed pre-registration (fingerprint c3e6841a206b8058, terminal CALORIE, stamped 2026-07-16T16:02:00Z):

PREREGISTERED 2026-08-12 shadow-projection (elements fp f9f9f8eebf7ed0cc): reykjavik lat=64146600 lon=-21942600 C1=60432 max=64126 C4=67658 obsmax=1000000 → PROJECT
latrabjarg lat=65502400 lon=-24525400 C1=60220 max=63934 C4=67495 obsmax=1000000 → PROJECT
coruna lat=43362300 lon=-8411500 C1=63056 max=66498 C4=69719 obsmax=1000000 → PROJECT
leon lat=42598700 lon=-5567100 C1=63164 max=66551 C4=69726 obsmax=1000000 → PROJECT
zaragoza lat=41648800 lon=-889100 C1=63281 max=66583 C4=68550 obsmax=1000000 → PROJECT
norwich lat=41524265 lon=-72075910 C1=61476 max=64520 C4=67443 obsmax=132007 → DEADCAT
millstone lat=42623300 lon=-71488200 C1=61246 max=64439 C4=67502 obsmax=158693 → DEADCAT
constants gate=800000ppm coupling=3/10 baseline=lower-median quiet=Kp<=4 bins=300s
resolution=post-eclipse crucible on measured vtec (2026-08-12)

Rendered verbatim from the sealed Form. invariant-verify --path …/eclipse-prediction-2026-08-12.invariant.json re-derives this text byte-for-byte from the Besselian elements.

In plain terms:

Station	Role	The claim
Reykjavik, Látrabjarg (Iceland) · A Coruña, León, Zaragoza (Spain)	totality path (obs = 10⁶ ppm)	each station-day box MUST PROJECT a whole child: TEC depletion ≥ 3/10 of obscuration inside its ≥80% gate window
Norwich CT (132,007 ppm) · Millstone Hill (158,693 ppm)	declared below-threshold controls	each MUST SEAL DEAD CAT — their obscuration can never open the 800,000 ppm gate, by geometry alone

The law is inherited, frozen, from the sealed 2024 benchmark — gate 800,000 ppm, coupling 3/10, lower-median baselines over Kp ≤ 4 control days, 300 s bins. The historical corpus's empirical coupling range (Historical corpus) is the expectation band; the 300‰ audit line is the pass/fail.

What falsifies it

1. Tampering: any change to the constants, the station roster, the coordinates, or the element set no longer re-derives the pinned text — invariant-verify --path …/eclipse-prediction-2026-08-12.invariant.json re-computes the geometry from the Besselian elements and byte-compares. Drift = broken pre-registration, visibly.
2. Physics: after the eclipse, a totality-path station with clean data that seals dead cat, or a declared control that projects, seals the resolution CURE — recorded raw, never renamed (LAW 3).
3. Storms: est Kp ≥ 5 on eclipse day does NOT void the prediction (the geometry side is untouched by a storm) — the resolution still runs; the interpretation carries a storm annotation and the control-channel claims weaken accordingly (pre-registered in Model shear).

The prediction EXTENSIONS — what depletion-magnitude models never publish

Born from the finished shear (Model shear): second-generation claims sealed on top of the base prediction —

1. The traveling signature: the five totality stations' depletion minima occur in a pinned ORDER with pinned gaps, each deepest bin riding its own station's max-obscuration second within the sealed lag window [−300, +1200] s (the bottomside delay measured −54…+1073 s across 2017/2023/2024) — the umbra's ground track written into the ionosphere.
2. The confinement claim (storm-proof): each totality station's excess in-window depletion clears a frozen threshold derived from the measured historical separation — decidable even if a G-storm hits on eclipse day, because a storm floods the flanks and a shadow cannot.
3. The quantified nulls: Norwich CT and Millstone Hill sub-gate depletion bounded numerically by the measured coupling band × their obscuration — a pre-registered undetectability claim.

Sealed prediction extensions (fingerprint b106d2926b1bf2cf, terminal CALORIE, stamped 2026-07-16T18:54:39Z):

PREREGISTERED-EXTENSIONS 2026-08-12 (elements fp f9f9f8eebf7ed0cc): TRAVELING depletion-minimum order latrabjarg@63934 +192s→reykjavik@64126 +2372s→coruna@66498 +53s→leon@66551 +32s→zaragoza@66583 each deepest bin riding its own max-obscuration second within [-300,+1200]s (the bottomside delay measured −54…+1073s across 2017/2023/2024, bin-quantized)
CONFINEMENT excess=window−flank(±7200s) median depletion ≥ threshold=50189ppm at every totality station AND < threshold at norwich+millstone (storm-proof: a storm floods the flanks and pins to window edges, a shadow cannot)
NULLS norwich obsmax=132007 expected-depletion=40790–73923ppm → UNDETECTABLE at 300s bins
millstone obsmax=158693 expected-depletion=49036–88868ppm → UNDETECTABLE at 300s bins
derived-from measured historical coupling 309–560permille
resolution=post-eclipse eclipse-shear analysis on measured 2026-08-12 vtec

Rendered verbatim from the sealed Form; invariant-verify --path …/eclipse-prediction-extensions-2026-08-12.invariant.json re-derives it byte-for-byte from the Besselian elements.

The resolution procedure (runs the day Madrigal posts 2026-08-12 — ~2026-08-30)

# 1. geometry for the 2026 stations is already ledger-stored; ingest the day + controls:
for d in 2026-08-12 2026-08-05 2026-08-19 2026-07-16; do
  gaiaftcl helio-ingest kp --day $d
  gaiaftcl helio-ingest madrigal-tec --day $d \
    --stations reykjavik,latrabjarg,coruna,leon,zaragoza,norwich,millstone
done
# (Madrigal posts on a measured two-to-three-week lag — eclipse day expected ~2026-08-30; see resolution status below.)

# 2. fossils for 2026 (quiet rule live), then the crucible against the sealed claim:
gaiaftcl eclipse-geometry --eclipse 2026-08-12 --build-fossils \
  --controls 2026-08-05,2026-08-19,2026-07-16
gaiaftcl eclipse-invariant --historical --seal --save   # the multi-eclipse run now includes 2026

# 3. wiki becomes the readout:
gaiaftcl eclipse-wiki-render

The verdict lands in the crucible tables on this wiki the moment the renderer runs — the same tables, the same law, the same fingerprints the world could copy down today.

Resolution status — measured T+1 (2026-08-13)

The sky kept the appointment; the archive has not yet. Measured on resolution day:

• Eclipse day was quiet. GFZ Kp for 2026-08-12: eight 3-hour bins, maximum Kp 2.667 — far under the Kp ≤ 4 quiet rule and the pre-registered est-Kp ≥ 5 storm gate. The storm flag never raised; the prediction grades at full strength, with no storm annotation. The flanking days measured quiet too: 2026-08-05 max Kp 1.667, 2026-08-11 max Kp 3.000.
• The response archive trails real time. Madrigal's world-gridded TEC (instrument 8000) has no experiment for 2026-08-12 yet — measured today, the archive's posting boundary sits between 2026-07-25 and 2026-08-01, a two-to-three-week lag. The sealed claims cannot be graded until the archive posts the day: grading on absent data would be a verdict about an empty query, not about the sky.
• The instrument is ready and unchanged. All four sealed Forms re-verified byte-for-byte today, 28 days after sealing; the crucible re-ran green (2024 benchmark 7/7, 0 false positives) and the shear discriminant re-derived its frozen threshold exactly (minimum eclipse excess 57,543 ppm vs storm floor 42,836 ppm — threshold 50,189 ppm standing).
• Expected grading window: the day Madrigal posts 2026-08-12 — at the measured lag, around 2026-08-30, by which time the third pre-registered control day (2026-08-19) posts alongside it, so the resolution runs exactly as sealed. A standing daily probe watches for the posting.

Nothing above changes a sealed claim; the fingerprints are untouched. A pre-registered experiment should say out loud when its grading data has not arrived — the absence of data is a measurement too.

Predictions vs first measured data — the raw-channel first look (added 2026-08-13, T+1)

The founder's requirement on resolution day was plain: the predictions and the measured data, side by side, now. The pinned grading archive has not posted eclipse day — but the raw receiver files captured and SHA-sealed this morning contain the ionosphere's answer, and the program computed it directly from those bytes on T+1. Method: GPS dual-frequency geometry-free phase (L1C/L2W) from the sealed RINEX files, sign-calibrated against the code combination, satellite arcs leveled, 300-second bins, differenced against the sealed control day 2026-08-05. This is relative slant TEC from single stations — a different observable from the pinned gridded surface the frozen law grades on. It is the first look, not the sealed verdict; the verdict still runs ~2026-08-30 on the pinned archive, untouched.

What the raw bytes say, station by sealed station:

Sealed station	Sealed maximum (27 days prior)	Measured deepest depletion (raw first look)	Lag vs sealed max	Depth vs control day	Sealed lag gate [−300, +1200] s
Reykjavik	17:48:46 UT	17:40–17:45 UT bin	−376 s (next-deepest bin −76 s)	−4.5 TECU	at the early edge; bracketing bins inside
A Coruña	18:28:18 UT	18:45 UT bin	+1,002 s	−6.2 TECU	inside
León	18:29:11 UT	18:45 UT bin	+952 s	−5.8 TECU	inside
Zaragoza	18:29:43 UT	18:35 UT bin	+467 s	−3.1 TECU (measured curve)	inside

Reading it plainly:

• Every one of the four sealed totality stations with a receiver in our custody measurably dipped, in its sealed window, on the day. Reykjavik's depletion is the textbook signature: onset ~17:30, bottom at 17:40–17:45 straddling the sealed 17:48:46 maximum, asymmetric recovery — the bottomside delay shape the historical corpus measured on three previous eclipses. The control day shows nothing at those minutes.
• The traveling structure is present at first-look resolution. Iceland dipped first; the Spanish trio dipped together roughly fifty minutes later — the sealed order. The sealed gaps within the Spanish trio (+53 s, +32 s) sit far below 300-second binning and single-station noise; they are for the pinned surface to grade, not this one.
• The depths are consistent with the sealed expectation band. Depletions of 3–6 TECU against evening quiet levels of roughly 8–15 TECU vertical at these latitudes sit inside the pre-registered 31–56%-of-quiet expectation — stated as consistency, not as a graded fraction, since arc-leveled relative TEC carries no absolute denominator.
• Recorded raw, not smoothed: A Coruña and León show a positive TEC excursion in the quarter-hour before their maxima (one León bin at +13 TECU — flagged as a possible arc-leveling artifact pending the pinned surface); Reykjavik's single deepest bin sits just outside the early edge of the sealed lag gate at this observable's resolution, with its neighboring bin inside; León's full-window deepest bin is a late feature at 19:15 near local sunset. These anomalies are printed here on the day they were measured, before the sealed grading, so nobody can accuse the record of meeting the data halfway.
• The null channel (Westford, beside Millstone Hill) is captured but not yet processed — its file is a different format generation; processing is queued. The sealed null claim is silence, and it grades on the pinned surface with everything else.

The derivation is in custody beside the raw files it read: the processing scripts, every per-station output, and a method README, all SHA-256-sealed in the same manifest. Anyone holding the RINEX files can re-run the arithmetic.

What this is and is not. It is the ionosphere's answer, read from bytes no one can alter, twenty-eight days after the claims were sealed and one day after the sky moved: four sealed stations, four measured dips, in the sealed windows, in the sealed order, at depths inside the sealed band. It is not the verdict — the verdict belongs to the frozen law running on the pinned archive, and that appointment stands at ~2026-08-30. If the pinned surface disagrees with this first look, both go on the page.

Current Form registry

Form	Kind	Class	Closure	Terminal	Fingerprint	Stamped
2024 benchmark	projection	shadow-projection	7/7	CALORIE	89afd825011f060e	2026-07-16T16:01:47Z
2026 prediction	prediction-2026-08-12	shadow-projection	PREREGISTERED	CALORIE	c3e6841a206b8058	2026-07-16T16:02:00Z
2026 prediction extensions	prediction-extensions-2026-08-12	shadow-projection	PREREGISTERED	CALORIE	b106d2926b1bf2cf	2026-07-16T18:54:39Z
historical universality	historical-universality	shadow-projection	16/20	CURE	cdf43e80f06f10a3	2026-08-18T16:42:21Z

Loaded and fingerprint-checked from ~/.gaiaftcl/franklin/invariants/ at render time.