URUANA — TWO-TEST AUDIT: LEGACY IGRF + MATCHED RESOLUTION
Date: 2026-08-14

TEST A — HISTORICAL IGRF GENERATION ROBUSTNESS
Question: Is the 1971-72 non-inversion of reconstructed total field an artifact of using a modern retrospective IGRF?

Source facts:
- CGBA acquisition: 1971-1972.
- PGW branch: IGRF-reduced historical digitization; WMC branch has unknown arbitrary MAG offset and is excluded.
- IGRF-1 coefficients: 1965 main field + 1965-75 secular variation (the model available during acquisition).
- IGRF-2: 1965 and 1975 main fields. In the archived coefficient files, the IGRF-2 1975 main coefficients are exactly the IGRF-1 1965 main plus 10 years of IGRF-1 secular variation, so IGRF-1 and IGRF-2 are identical for 1971-72 under linear interpolation.

Results at Uruana / 0.79 km altitude:
- IGRF-1/IGRF-2 main-field northward gradient: +1.444 to +1.459 nT/km across 1971.0-1972.5.
- Current retrospective IGRF14/DGRF gradient: +1.354 to +1.380 nT/km.
- The sign is positive under every model/date.
- Adding the observed PGW crustal residual slopes (-0.233 to -0.356 nT/km) gives reconstructed total gradients:
  legacy IGRF-1/2: +1.093 to +1.227 nT/km;
  retrospective IGRF14: +1.006 to +1.147 nT/km.
- 0 reconstructed cases invert.
VERDICT A: PASS for robustness of the 1971-72 total-field non-inversion. It is not caused by choosing a modern IGRF.

TEST B — MATCHED-RESOLUTION / DEGRADATION HYPOTHESIS
Question: If 2008/2011 are blurred to the effective sampling of the historical CGBA, does the 2011 total-gradient inversion disappear while 1DV remains?

Historical sampling around Uruana:
- CGBA E-W line spacing median ~2.085 km (local range ~1.50-2.45 km).
- MAG values are strongly contour/digitization-like; ~81.8% of local values are exact multiples of 2.5 nT.

2011 total-field emulation on the same frozen observed north cap (6-8 km north within r<=8 km):
- Frozen primary result: 14/14 raw lines inverted; median raw slope -1.649644 nT/km; median IGRF +0.998354 nT/km.
- 2-km FWHM low-pass of raw-minus-IGRF residual, then add IGRF: 14/14 inverted; median -1.067 nT/km.
- Same 2-km low-pass + 2.5-nT residual quantization: 14/14 inverted; median -1.180 nT/km.
- Extreme 6-km low-pass + 2.5-nT quantization: 11/14 inverted; median -0.739 nT/km.
Thus a simple 2-km blur/quantization does NOT erase the 2011 total-gradient inversion.

2008 total-field N-S gradient inside frozen r<=8 km:
- native raw grid: +0.151 nT/km;
- 2-km FWHM: +0.164 nT/km;
- 6-km FWHM: +0.072 nT/km.
Thus 2008 did not show the 2011-style total-field N-S inversion even before degradation.

1DV behavior:
- 1971-72 CGBA: all 450 reconstruction combinations retain a 1DV zero within 8 km; this includes 90/90 at 6-km upward continuation.
- 2008: prior frozen scale-persistence audit already retained the inversion through 0-6 km continuation across raw/corrected x detrend 1-3 families.
- 2011 numeric 1DV: nearest zero ~6.04 km native; after 2-km Gaussian low-pass it remains ~7.02 km, still inside 8 km. At 4-6 km the edge-limited valid support collapses, so no stronger inference is made.

VERDICT B: FAIL for the simple 'resolution alone explains the 1971 total-field FAIL' hypothesis. At the actual ~2-km CGBA-like scale, the 2011 total-gradient inversion remains 14/14 even with 2.5-nT quantization. The more stable cross-campaign observable is the crustal derivative/sign-boundary architecture, not total-field inversion.

NEW SYNTHESIS — GRADIENT COMPETITION
Define a descriptive detectability ratio R = |dF_crust/dN| / (dF_main/dN). This is NOT a temporal-strength statistic because survey support/resolution differ.
- 1971-72: R ~0.17-0.26; total stays positive.
- 2008: R ~0.86; total stays positive.
- 2011 observed cap: R ~2.63; total flips negative.
The threshold R>1 naturally explains when a crustal gradient opposed to the main field becomes visible as an inversion of the total scalar-field gradient.

INTERPRETATION BOUNDARY
Do not claim crustal magnetism physically strengthened from 1971 to 2011: acquisition geometry, resolution, map digitization, support, altitude, and processing differ. The robust claim is that an opposed crustal-gradient/1DV boundary recurs independently; only the 2011 observed cap demonstrates direct total-field gradient reversal.