Eclipse 2026 Science Research

The science — August 12, 2026, and everything history says about it

This page is the research foundation of the experiment: what the 2026 eclipse is, what measured ionospheric physics says a total eclipse does, what the reference storms look like, and what the solar cycle does to the noise floor. Data tables inside GAIA markers are generated from the substrate's own ledger.

---

1. The August 12, 2026 total solar eclipse

Global circumstances (NASA/GSFC Espenak lineage; the substrate's embedded Besselian elements carry fingerprint f9f9f8eebf7ed0cc):

• Instant of greatest eclipse: 2026 Aug 12, 17:45:53.8 UT (17:47:06 TD; ΔT = 71.4 s) at 65°13.5′N 025°13.7′W — the North Atlantic ~45 km off western Iceland.
• Greatest duration: 2 min 18.2 s near 65°50′N; Sun altitude at greatest eclipse only +25.8°, azimuth 248.4°; path width 294 km.
• Magnitude 1.0386 · gamma 0.8977 · Saros 126 (member 48 of 72, descending node).
• Global contacts: partial begins 15:34:11 UT → total begins 16:58:05 UT → greatest 17:45:54 UT → total ends 18:34:05 UT → partial ends 19:57:56 UT.
• The path: begins at sunrise over the Taymyr Peninsula, sweeps (east-to-west in its first half — unusual) across the Arctic Ocean near the North Pole (~17:06 UT), clips eastern Greenland (Scoresby Sund ~17:15 UT), crosses western Iceland (Westfjords, Snæfellsnes, Reykjanes), the North Atlantic, and northern Spain (A Coruña, León, Bilbao, Zaragoza, Valencia, Palma) near sunset. Madrid and Barcelona sit just outside totality.
• Historical rarity: the first total eclipse visible from Iceland since June 30, 1954 and from Reykjavík since June 17, 1433; the first total eclipse over mainland Spain since August 30, 1905.

The substrate's own station circumstances for 2026 (computed by eclipse-geometry, stored in the ledger):

Station	C1 (UT)	Maximum (UT)	C4 (UT)	Max obscuration
albuquerque	—	—	—	0 ppm (0.0%)
carbondale	—	—	—	0 ppm (0.0%)
casper-wy	—	—	—	0 ppm (0.0%)
cleveland	17:07:12	17:40:00	18:12:16	29458 ppm (2.9%)
columbia-sc	—	—	—	0 ppm (0.0%)
corpus-christi	—	—	—	0 ppm (0.0%)
coruna	17:30:56	18:28:18	19:21:59	1000000 ppm (100.0%)
dallas	—	—	—	0 ppm (0.0%)
eugene-or	—	—	—	0 ppm (0.0%)
indianapolis	—	—	—	0 ppm (0.0%)
latrabjarg	16:43:40	17:45:34	18:44:55	1000000 ppm (100.0%)
leon	17:32:44	18:29:11	19:22:06	1000000 ppm (100.0%)
mazatlan	—	—	—	0 ppm (0.0%)
millstone	17:00:46	17:53:59	18:45:02	158693 ppm (15.9%)
nashville	—	—	—	0 ppm (0.0%)
norwich	17:04:36	17:55:20	18:44:03	132007 ppm (13.2%)
reykjavik	16:47:12	17:48:46	18:47:38	1000000 ppm (100.0%)
salem-or	—	—	—	0 ppm (0.0%)
san-antonio	—	—	—	0 ppm (0.0%)
zaragoza	17:34:41	18:29:43	19:02:30	1000000 ppm (100.0%)

Besselian elements fingerprint f9f9f8eebf7ed0cc · quantized 10-s lattice · generated from helio_eclipse_circumstances.

Norwich, Connecticut — the home station — sees a shallow ~13.2% partial with the Sun ~60° high at maximum 17:55:20 UT (1:55 PM EDT). That is a fundamentally different physical regime from totality: far too weak for a detectable local TEC signature, so Norwich (and Millstone Hill, 15.9%) are the declared below-threshold control channel of the 2026 prediction — a falsifiable claim in its own right. The science payload comes from the totality path: Iceland and Spain.

Path geometry caveat: 2026 is a high-latitude, low-Sun (+25.8°), near-sunset-over-Europe event; the 2017/2024 references are mid-latitude, high-Sun US events. Quantitative transfer of depletion magnitudes is approximate — which is exactly why the pre-registered law audits a relation (depletion commensurate with obscuration), not an absolute number.

2. What a total eclipse measurably does to the ionosphere

The Moon's shadow switches off the ionosphere's photoionization source along a supersonic (~700 m/s at F-region heights) moving track. Measured consequences from the two best-instrumented modern eclipses:

Total Electron Content (TEC) depletion

• 2017-08-21: TEC decreased 30–40% within the 75% obscuration area with equatorward expansion (Cherniak & Zakharenkova, GRL 2018); peak depletions ~50–60% vs background (Coster et al. 2017); tomography showed ~40% electron-density depletion vs the prior day, recovering ~30 min after totality (EPS 2022); COSMIC radio-occultation profiles ~40% maximum depletion.
• 2024-04-08: reductions 20–65% depending on obscuration and distance from totality. Millstone Hill ISR (Aa et al., JGR Space Physics 2024): above 250 km, Ne dropped ~60% and persisted up to 6 hours; below 250 km, <30% lasting 2–3 h. SAMI3-TIDAS assimilation: Ne reduction 20–50%, max in the F-region 200–250 km, lag growing with altitude (5–10 min bottomside → 1–2.5 h topside). Absolute depletions ~8–12 TECU after non-uniform background correction (GPS Solutions 2025). 2024 also produced the first reported post-eclipse TEC enhancement propagating equatorward, followed by a secondary depletion belt.

Waves

• Eclipse-induced ionospheric bow waves (Zhang et al. 2017); gravity waves with 25–90 min periods, peaking 25–30 min, in 30-s GPS-TEC (Nayak & Yiğit 2018); modeled supersonic-shadow bow-wave front + mesoscale acoustic-gravity waves (Lin et al., GRL 2018); eclipse-driven medium-scale TIDs documented for 2024 (Liang et al., Space Weather 2026).
• Neutral-atmosphere/infrasound detection is historically marginal — the 2015 UK study (Marlton et al.) could not confidently attribute surface pressure/temperature wave perturbations to the eclipse; treat any such channel as exploratory.

Timing structure — peak TEC/foF2 reduction lands within ~30 min of local maximum obscuration; the topside lags by 1–2.5 h; foF2 overshoots on recovery. The substrate's projected leaves (below) land within minutes of local maximum — consistent with the published bottomside lag.

Raw vs smoothed products: high-cadence line-of-sight/1 Hz GNSS data preserves the bow waves and short-period gravity waves that gridded, binned vertical-TEC maps average away, and the choice of background-correction method alone moves reported depletion magnitudes by 20–40%. Engineering consequence baked into the experiment: ingest raw-est available products, quantize once, pin the baseline rule (lower median over named quiet days) in the sealed Form so the correction method can never drift post-hoc.

3. What the substrate measured (the historical corpus)

The corpus below is the substrate's own end-to-end measurement — public raw archives → quantized integer ledger → the ONE frozen pair law. Every row is generated from the live crucible:

Eclipse	Station	Projected (UT)	Obscuration (ppm)	Depletion (ppm)	Coupling (‰ of obs)	Day max Kp
2017-08-21	carbondale	18:05:47 UT	800292	391086	488	3.0
2017-08-21	casper-wy	17:29:17 UT	802380	268279	334	3.0
2017-08-21	columbia-sc	18:27:38 UT	800290	448461	560	3.0
2017-08-21	nashville	18:12:52 UT	802211	392061	488	3.0
2017-08-21	salem-or	17:05:17 UT	802393	263023	327	3.0
2023-10-14	albuquerque	16:26:33 UT	801694	265351	330	3.0
2023-10-14	corpus-christi	16:46:38 UT	800667	315295	393	3.0
2023-10-14	san-antonio	16:42:59 UT	801810	319424	398	3.0
2024-04-08	carbondale	18:46:29 UT	800534	274457	342	3.3
2024-04-08	cleveland	19:01:31 UT	802627	353317	440	3.3
2024-04-08	dallas	18:45:00 UT	997624	308350	309	3.3
2024-04-08	indianapolis	18:53:22 UT	800050	286986	358	3.3
2024-04-08	mazatlan	18:05:00 UT	970682	323319	333	3.3
2024-04-08	millstone	19:17:05 UT	801464	419091	522	3.3
2024-04-08	norwich	19:17:06 UT	801484	426439	532	3.3
2026-08-12	latrabjarg	17:40:00 UT	915557	293964	321	2.7

Coupling = depletion·1000 ÷ obscuration at the projected second (exact integer). The 2026 expectation inherits this measured range: at totality (obs = 10⁶ ppm) expect depletion between the table's min and max coupling ‰ of baseline; the frozen law audits at ≥ 3·obs/10.

See Historical corpus for the full per-eclipse picture and Benchmark results for the sealed 2024 verdict.

4. The reference storms (the data that shears naive models)

A geomagnetic storm moves TEC as violently as an eclipse — without a shadow. Any model that detects eclipses by "TEC dropped" false-positives on storm days. The two canonical references:

• May 10–11, 2024 "Gannon" superstorm: minimum Dst −412 nT — the most intense G5 storm since March 1989; Kp reached 9 twice and stayed ≥8 for 24 hours; daily Ap 273; magnetosphere compressed to ~5 R_E; aurora to Texas, Florida, Spain, Italy. Driven by successive CMEs from AR13664 (incl. an X5.8 flare). Magnitude a 1-in-12.5-year event, duration 1-in-41-year (Elvidge & Themens 2025).
• October–November 2003 "Halloween" storms: Dst −383 nT (Oct 30) and −422 nT (Nov 20); dayside TEC increases of ~40% (Oct 29) and ~250% (Oct 30) within hours via prompt-penetration electric fields — the "daytime super-fountain" (~900% electron-content increase above CHAMP altitude); the Oct 28 X17 flare alone stepped TEC ~30% (~25 TECU) in ~15 min.

Both storms' TEC days are ingested into the ledger; the Gannon days form the shear-negative fossils — see Model shear. The 2026 operational rule is pre-registered: max Kp > 4 disqualifies a day from any baseline, and eclipse-day est-Kp ≥ 5 flags the storm pivot (the path-station resolution still runs; the interpretation gains a storm annotation and the storm archives become the science channel).

5. Solar-cycle context — why history spans three regimes

• 2017-08-21 sat in the declining phase of Solar Cycle 24: low background TEC (roughly half the 2024 levels), quiet variability — the low-noise regime.
• 2023-10-14 (annular) sat on rising SC25.
• 2024-04-08 sat near SC25 maximum.
• 2026-08-12 sits near/just past SC25 maximum: per NOAA SWPC (Jan 30, 2026), SC25 likely reached the highest sunspot number in over 20 years — WDC-SILSO's initial daily estimate for Aug 8 was 299 (SWPC non-official estimate 337, a value unseen since March 2001). SC25 ran ~31% above SC24 at the equivalent point.

Consequences: 2026 background TEC and day-to-day variability resemble 2024, not 2017 — the noise floor is high, storm probability was elevated (August 12 itself resolved geomagnetically quiet), and the cross-cycle corpus (same law, three regimes) is precisely what bounds how much of the eclipse signature survives regime change.

6. Live context

• Latest SWPC estimated Kp in ledger: 1.33 at 2026-08-18 11:36:00 UT
• Latest GFZ definitive/nowcast Kp day in ledger: 2026-08-12
• Latest magnetometer capture: BOU H at 2026-08-18 11:36:00 UT

Ledger-borne (renders identically until the next ingest).

Source register

NASA/GSFC Five-Millennium-Canon Besselian elements (2017/2023/2024/2026, embedded verbatim, fingerprinted) · nationaleclipse.com + eclipse2024.org city circumstances (fixture anchors: Dallas, Norwich CT, Carbondale 2017, Albuquerque 2023) · Cherniak & Zakharenkova GRL 2018 · Coster et al. 2017 · Zhang et al. 2017 · Nayak & Yiğit 2018 · Lin et al. GRL 2018 · Aa et al. JGR Space Physics 2024 · GPS Solutions 2025 (background correction) · Liang et al. Space Weather 2026 · Marlton et al. (2015 UK) · Elvidge & Themens 2025 (Gannon statistics) · NASA NTRS 20070019835 (Halloween TEC) · NOAA SWPC news 2026-01-30 (SC25 record SSN) · MIT Haystack Madrigal / GFZ Potsdam / USGS / NOAA SWPC archives (live-verified — see Data archives).