URUANA 2026 — NOAA WMM/WMMHR/IGRF TOTAL-FIELD GRID AUDIT

INPUT FILES
- WMMHR-2025, date 2026.61644, elevation 0.700 km
- WMM-2025, date 2026.61644, elevation 0.700 km
- IGRF2025, date 2026.61644, elevation 0.700 km
- 16 nodes in each file (4 latitudes x 4 longitudes)

IMPORTANT GRID LIMIT
Actual node spacing in the uploaded output is 0.10000 degree in latitude and longitude, not 0.01 degree.
Near Uruana this is ~11.1 km N-S and ~10.7 km E-W.
The city coordinate (-16.0593,-46.2585) is not itself a node.
Closest node: (-16.0393,-46.2485), distance ~2.47 km.

PRIMARY PREDECLARED PROXY
DeltaF_HR = F_WMMHR - F_WMM.

CENTRAL LONGITUDE TRACK (lon=-46.2485, nearest to city)
WMM north-gradient: 1.139186 nT/km
WMMHR north-gradient: 1.308354 nT/km
IGRF north-gradient: 1.142785 nT/km
WMMHR-WMM increment north-gradient: 0.169168 nT/km
Signed increment/main ratio: 0.148499

2-D PLANE FIT OVER ALL 16 NODES
WMM dF/dN: 1.139186 nT/km
WMMHR dF/dN: 1.292382 nT/km
IGRF dF/dN: 1.142110 nT/km
WMMHR-WMM increment dF/dN: 0.153196 nT/km

RESULT
- WMM and IGRF both predict positive northward total-field gradient.
- WMMHR also predicts positive northward total-field gradient.
- The WMMHR-WMM increment is also positive northward at every sampled longitude.
- Therefore the 2026 WMMHR total-field model DOES NOT reproduce the negative N-S total-field-gradient inversion observed in the 2011 airborne data.
- At this model scale the high-resolution increment reinforces, rather than opposes, the positive main-field gradient.

STATUS
FAIL for current total-field inversion in WMMHR2025.
This is a model result, not a 2026 local measurement.
Because WMMHR spatial resolution is far coarser than 8 km, this FAIL does not falsify persistence of a local crustal anomaly below model resolution.

NEXT TEST
Use Z (vertical component) for WMM, WMMHR, and IGRF. This directly tests whether the high-resolution model contains a vertical-component contribution of opposite sign even though F does not invert.
For a materially higher-resolution modern model, HDGM2026 is the more appropriate model-scale follow-up.