Historical Eclipse Validation — Pure-Tidal ΔT vs the Documented Record

This page is the empirical confirmation that the model’s predictions actually hold against observation. For the closed-form ΔT formula itself, see Timekeeping & Delta-T; for the Moon polynomial’s implementation and parallax limit at modern epochs, see Sun, Moon & Planets; for the broader validation tradition (Wells 1963, Cheng 2016, etc.), see Supporting Evidence.

1. Thesis

Pure-tidal Farhat-based Moon orbital evolution and the resulting secular ΔT are sufficient to explain documented solar-eclipse visibility across at least 2,400 years of historical records.

The conventional interpretation of the long-term ΔT record holds that pure tidal physics alone cannot account for the observed eclipse data — that an additional non-tidal Earth-rotation component is required, attributed to glacial isostatic adjustment and core-mantle coupling (the Munk-MacDonald hypothesis). Stephenson’s empirical polynomial reproduces the observations and is therefore consistent with this non-tidal component being present.

The model’s claim is the opposite: tidal physics alone, derived from the Farhat 2022 deep-time evolution model and applied to the Earth-Moon system via angular momentum conservation, already fits the eclipse record at least as well as Stephenson — and on the clean visibility-window methodology, slightly better.

This does not prove non-tidal speedup is absent. The eclipse data is noisy enough that a small non-tidal component could be present in Earth’s actual rotation history without leaving a detectable signature in the visibility-window comparison. What the validation establishes is that pure-tidal is not falsified by the eclipse record. The conventional view that non-tidal speedup is required turns out to be an over-claim.

2. Three independent validations

The validation runs in three layers, each independently informative.

Layer 1 — Moon polynomial vs NASA in TT space (foundation)

The simulation’s Moon position uses Meeus Ch. 47 (60 longitude + 60 latitude perturbation terms) on top of a 5-layer geometric precession hierarchy. For 11 canonical eclipses from NASA’s Five Millennium Catalog of Solar Eclipses spanning -524 to 985 CE, the model’s computed time of Moon-Sun conjunction was compared to NASA’s published Terrestrial Dynamical Time (TD) of greatest eclipse.

The comparison is ΔT-independent: in TT-space, the astronomical event is fixed regardless of which ΔT either side assumes. Any residual is purely a Moon polynomial accuracy question.

This is the expected Meeus Ch. 47 polynomial residual at these timescales (≈ 0.13° in Moon ecliptic longitude at year 980, the worst case). The polynomial is sound at every epoch tested.

Layer 2 — Per-event same-day conjunction check

For 19 documented historical solar eclipses with known calendar date and observation site, the model’s Moon-Sun ecliptic longitude separation Δλ was evaluated at noon UT on the documented date. The question: is there a conjunction within ±12 hours of that noon (same calendar day)?

Result: 19/19 same-day matches. Every documented eclipse date corresponds to a conjunction in the model.

Among the 19, the sub-solar-distance breakdown is:

The single ✗ event is the Chinese Spring/Autumn observation (-708, Lu State at 113°E). This is a test-method artifact rather than a model failure: the noon-UT proxy used by Layer 2 evaluates sub-solar position at 12:00 UT, but at Lu State’s longitude the local noon falls in the early UT morning, so the 12:00 UT sub-solar lands 7.5 hours past local noon and far west of the site. The Layer 3 visibility-window test (which scans ΔT to find the actual eclipse time, not just noon UT) confirms the Chinese event is in the model’s visibility region — see §Layer 3 below.

Layer 3 — Visibility window: pure-tidal vs Stephenson (the headline)

For each event, the question becomes: what range of ΔT values would put the model’s eclipse path within penumbra reach (< 7,500 km) of the observation site? And do either of (a) the model’s pure-tidal ΔT or (b) Stephenson’s empirical ΔT fall inside that range?

Pure-tidal wins both the visibility count and the mean-residual test. The penumbra count is the headline: if Stephenson’s empirical curve were “the truth”, it should clearly win the broad visibility test — Stephenson’s coefficients were calibrated to make eclipses visible at observed sites. Instead, pure-tidal explains more of the visibility, not less.

The two events where Stephenson loses penumbra are Ibn Yunus 979 May 28 and 1004 Jan 24 — both fail because Stephenson’s ΔT is too low at those medieval epochs, pushing sub-solar east of Cairo. The model’s higher pure-tidal ΔT recovers them.

The umbra-count tie at 6/13 each is the noise floor — totality strips are narrow (~270 km wide) and ancient localization is too coarse to discriminate at sub-100 s ΔT precision.

3. The ΔT gap — two readings

The model’s pure-tidal ΔT is consistently higher than Stephenson’s empirical fit, by an amount that scales linearly with time:

The linear-in-time slope of ~1.96 s/yr corresponds geometrically to a constant Length-of-Day difference of ~5–6 ms between pure-tidal and Stephenson — the order of magnitude of canonical non-tidal Earth-rotation estimates (the Munk-MacDonald mechanism, conventionally attributed to glacial isostatic adjustment plus core-mantle coupling).

Two equally consistent readings of this gap exist:

• Mainstream reading: Earth has a real non-tidal Earth-rotation component of the magnitude estimated by Munk-MacDonald; Stephenson’s empirical curve captures it; pure-tidal alone is incomplete by that amount.
• Model reading: Stephenson’s polynomial is an empirical fit to the eclipse record, not an independent measurement of a non-tidal component. The mainstream attribution of the polynomial-vs-pure-tidal gap to a real non-tidal mechanism is one interpretive choice; the gap is equally consistent with pure-tidal Farhat physics already explaining the record at the noise floor of historical observation precision.

The historical eclipse data alone cannot distinguish these two readings. Both give the same observed eclipses. The choice rests on:

• How much weight is placed on independent Earth-rotation evidence (geological cyclostratigraphy, modern Lunar Laser Ranging)
• Whether parsimony — a single tidal Q-factor as the only free parameter for Earth-rotation history — is preferred over a phenomenological two-component model
• The Hadean self-validation at the Moon-at-Roche limit that the pure-tidal model passes without using any Hadean data in the fit (see Expanding Resonance)

4. Methodological corrections

The clean visibility-window comparison above was only possible after three methodological corrections to the prior pipeline. These are documented in detail in the simulation repo’s doc 101 ; summarised here:

Sub-solar sign-bug fix — The previous code applied ΔT in the wrong direction when computing sub-solar geographic location, shifting all ancient eclipses ~97° west of where they should appear. After the fix, the mean sub-solar offset across the 19 events dropped from ~8,350 km to ~4,700 km.

Gregorian / Julian auto-switch — Calendar dates after 1582-10-15 are now correctly interpreted as Gregorian (matching the convention used by Wikipedia and NASA’s Five Millennium Catalog for naming historic eclipses). Pre-1582 dates remain proleptic Julian.

Mis-attributed test entries removed — One entry in the prior pipeline was a category error: “Cambyses (-522)” had been labelled as a solar eclipse, but Stephenson 1997 confirms no documented Babylonian solar eclipse exists from this era. The Cambyses-era eclipses referenced in the Babylonian astronomical diaries (16 Jul 523 BC and 10 Jan 522 BC) are both lunar. With Cambyses removed, the cleaned test set is 19 events.

After these corrections, the validation became a clean apples-to-apples comparison of two ΔT models on geometrically correct eclipse paths.

5. What the validation establishes

What the model can claim with confidence:

1. The Moon polynomial used in the simulation is validated against NASA’s JPL reference at ±15 min back to 2,500 years before J2000 (n = 11 events, -524 to 985 CE).
2. The model’s pure-tidal ΔT formula explains documented solar-eclipse visibility for 19/19 events spanning -762 to 1654 CE.
3. No falsifying counterexample exists in the cleaned 19-event record that requires invoking a non-tidal Earth-rotation component.
4. The pure-tidal model’s deep-time grounding is independent of the eclipse record — anchored to Wells 1963 (Devonian coral growth bands), Wu et al. 2024 (650-Myr cyclostratigraphy), and modern Lunar Laser Ranging. None of this evidence is circular with the historical eclipses.

What the model is not claiming:

• That Stephenson’s curve is wrong. It fits the eclipses well too.
• That non-tidal speedup is definitively absent. The data is noisy enough that a small non-tidal component could be present but not discriminated from the pure-tidal null at this resolution.
• That solar-eclipse data alone settles this. Lunar-eclipse timing — geographic-noise-independent — and high-precision LOD reconstructions are stronger constraints.

6. Limits

Caveats the reader should keep in mind:

1. n = 19 historical solar eclipses is small. Statistical power for discriminating models with sub-100 s ΔT differences is limited.
2. Geographic localization of ancient eclipse paths has irreducible uncertainty. The 4,500 km umbra reach and 7,500 km penumbra reach are approximations based on the model’s sub-solar-point distance.
3. Ancient observation sites often have lat-err > umbra reach (e.g., 977 Dec 13: Cairo is 3,000 km north of the umbra path regardless of ΔT). For these events the test only tells us about penumbra visibility, not totality.
4. The visibility-window methodology cannot detect ΔT differences smaller than ~50 s because the penumbra window is wide (~30,000 s typical). For finer ΔT discrimination, lunar-eclipse timing-at-site is the stronger constraint.

7. Open question — the Thales eclipse date

While exhaustively searching for Thales-era candidates in the simulation, an alternative date emerged from the model that does not appear in the standard scholarly candidate list:

Herodotus 1.74 describes the Thales eclipse as a dramatic darkening of the sky that halted the Battle of Halys between the Lydians and the Medes. A penumbra-only partial of magnitude ~0.5 (which is what the traditional -584 May 28 date gives in the model) is geometrically inconsistent with that account.

The -582 March 28 candidate places the umbra centerline directly across the conflict zone — the physical geometry Herodotus’s account requires.

This finding is suggestive but not definitive. A path passing over Anatolia doesn’t automatically mean this event was the Thales eclipse; an eclipse a few years away from the traditionally-accepted year could pass through the same region by coincidence. The validation deliberately keeps the traditional -584 May 28 date in the test data pending an independent scholarly review of the -582 March 28 candidate.

8. Reproducing the validation

The complete validation suite is available in the simulation as developer-mode console tests at Console Tests (F12) > Historical Eclipses & ΔT (10 buttons): the NASA catalog cross-check (Layer 1 above), the per-event same-day check (Layer 2), the visibility-window comparison (Layer 3), plus 7 supporting diagnostics. The 19 documented eclipses plus 8 modern reference events are also exposed as planetStats nav-buttons under Moon → Historical Solar Eclipses (validation), where the user can step through each event and visually verify Moon-Sun alignment in the 3D scene.

Full reproducibility notes, methodology, and underlying numerical inputs are in the simulation repo’s doc 101 . The doc 100 baseline (35-eclipse residual-RMS comparison from 2026-06-18, which preceded the sub-solar bug fix and Cambyses cleanup) is preserved at doc 100  as the prior reference.

Continue to Predictions for the model’s testable predictions across near-term, medium-term, and deep-time horizons — including the lunar-eclipse follow-up that would further tighten this comparison.