GeaSpirit
Physical evidence, remote sensing & subsurface R&D
GeaSpirit tests physical and remote-sensing signals as possible sources of geological context. These research lines are experimental, may return negative results, and do not affect Evidence Compass unless they are independently validated.
Two reference chapters on what remote and geophysical methods actually measure — written to keep measurement, inference and discovery apart. They describe methods, not GeaSpirit capabilities; where something is a GeaSpirit idea, it is labelled as a proposal.
Gravity responds to variations in mass and density beneath the surface. GeaSpirit uses it as contextual physical evidence to identify large-scale density contrasts and geological structures that may be relevant to mineral systems. A gravity anomaly is not a mineral discovery.
These two numbers are not the same thing, and the map shows the smaller one. 322,954 assets have a gravity and magnetic reading — that is the coverage. 11,966 of them additionally carry an anomaly summary, which is what a marker on this map represents. An asset without a marker is not an asset without data: it has a reading whose local variation was not distinct enough to summarise. Absence of a dot is not absence of gravity, and it is certainly not negative evidence.
Residual gravity, gravity gradient and multiscale context are PLANNED. Not built as an active analytical layer. Tested across 12 known mineral districts using global Level 1 gravity data; no detectable mineral-scale signal was observed at the spatial resolution supported by the source dataset. The measurement was repeated with a full terrain correction and the result did not change.
Loading gravity context…
If the rock beneath a place is denser than its surroundings, local gravity is very slightly stronger. If it is less dense, very slightly weaker. That is the whole physical basis, and it is why gravity is one of the few remote signals that answers to material below the surface rather than to the surface itself. But the relationship runs one way only: a denser body produces an anomaly, and an anomaly does not identify a body.
Large intrusions. Basin margins. Basement steps. Regional density boundaries. Large crustal structures. The architecture a mineral system sits inside — and often the structure that controls mineralisation is more detectable than the mineralisation itself.
Global gravity cannot directly identify copper, gold, REE or massive sulphide, and it states no resource or reserve. A body 100–1,000 m across is below the effective resolution of this dataset at any depth — not because it is deep, but because the data does not carry that scale.
The global grid is served at about one arc-minute, and it is tempting to read that as roughly two-kilometre detail. It is not. Over the ocean the grid really is derived from satellite radar altimetry of sea-surface height; over land there is no altimetric gravity, and the same distributed grid carries values from a global model. Both are served at the same step, so at the point of use they look identical. GeaSpirit measured the difference on equal windows instead of assuming it:
| Window | Std. dev. | Decorrelates at |
|---|---|---|
| Ocean — central Atlantic | 15.1 mGal | 32.0 km |
| Ocean — central Pacific | 4.4 mGal | 23.8 km |
| Land — Nevada, USA | 51.3 mGal | 7.2 km |
| Land — Western Australia | 14.5 mGal | 13.3 km |
| Land — Atacama, Chile | 71.0 mGal | 11.9 km |
Land is not smoother than ocean — it carries more short-wavelength structure, not less. But 7–13 km is where the land field stops being self-similar, and that is the smallest thing it can resolve. Every scale claim on this page uses that measured figure rather than the served step.
Free-air anomaly is not corrected for the mass of the terrain, so it could plausibly be an elevation ranking wearing a geophysical name. Measured against 300 land points whose elevation came from a source never derived from gravity, elevation explained about 9% of the variance (r = 0.302), and the fitted slope was 6% of the textbook free-air correction. Isostatic compensation is why: the root of lighter crust under a range largely cancels the topographic excess.
That 9% is a GLOBAL figure, and it means nothing without its scale. Measured inside a single 100 km window across twelve mining districts, the same coupling reads r ≈ 0.72 — about half the variance; between windows it falls to 0.52, and across the planet to 0.30. The reason is that isostasy is a long-wavelength phenomenon: a mountain belt has a crustal root that cancels its mass, but a single ridge does not. So free-air is terrain-dominated locally and terrain-free regionally, and any statement about the coupling has to say at which scale it holds.
That result matters because it changes the argument. The reason to move towards Bouguer and residual products is not that free-air “just measures mountains” — it is that free-air is insensitive to compensated relief, and therefore blind to exactly the density structure this layer is meant to expose. These figures are a GeaSpirit internal audit on a 300-point sample, not a universal physical constant.
Priority and confidence are not the same quantity. A mass anomaly can be real, well measured and clearly present while its geological meaning stays uncertain — for example mass anomaly: HIGH alongside interpretation confidence: LOW. That pair is a valid and common outcome, and reporting it honestly is more useful than collapsing the two into a single number that implies more than the physics supports.
Gravity responds primarily to density; magnetics responds to magnetic properties. They are independent measurements, so where they agree the structural reading is better supported — but convergence identifies structure, not mineralisation. The combinations below are conceptual guidance, never a diagnosis.
Compatible with several dense and magnetic geological bodies — mafic or ultramafic rock, an intrusion, magnetite-bearing material. Compatible with, not evidence of.
A dense but weakly magnetic source is one possibility among several. Requires geological context before it means anything.
May indicate a basin, a less dense granite, alteration or fracturing. May — the list is not exhaustive and the cases are not exclusive.
Where a gravity edge and a magnetic edge line up, that may indicate a geological contact, a fault or the margin of an intrusion. Structure is often more detectable than the deposit it controls.
| Level | Source | Useful for | Status |
|---|---|---|---|
| Level 1 — Global | Compiled global grids | Regional context, structure at roughly 10 km and larger | BUILT |
| Level 2 — National | Geological survey gravity | Regional exploration structure; station spacing depends on the survey | NOT BUILT |
| Level 3 — Asset | Ground, airborne, gradiometry | Asset-scale investigation where such data exists | NOT BUILT |
Levels 2 and 3 are a data-acquisition problem, not a processing one. No residual, gradient or multiscale filter applied to the global grid can recover detail that was never in it — filtering redistributes what is present and invents nothing. Those products are therefore listed as planned rather than offered as layers.
Density contrast and regional structure
Nocturnal thermal memory research
Gamma-ray K, Th and U against geological expectation
Surface mineralogy against geological expectation
Multiscale similarity to known deposit families
Local similarity to individual systems, not to their average
…
…
Reopen criteria …
…
Reopen criteria …
…
Reopen criteria …
…
Reopen criteria …
| Line | Last experiment | Result | Why | What would reopen it |
|---|---|---|---|---|
| … | T3 · Level 2 | … | … | … |
| … | N1.4 | … | … | … |
| … | R1 | … | … | … |
| … | S1 gate | … | … | … |
…
…
| Research line | Status | What happened | What would reopen it |
|---|---|---|---|
| … | … | … | … |
| … | … | … | … |
| … | … | … | … |
| … | … | … | … |
| … | … | … | … |
| … | … | … | … |
| … | … | … | |
| … | … | … | |
| … | … | … | |
…
Gravity–magnetic fusion and depth hypotheses are research, not a service, and they live in GeaSpirit Private Intelligence, in the Depth Fusion Lab. Their status there is stated plainly: the magnetic inversion has only ever been tested against synthetic truth, never against real survey grids, and the depth estimator failed its own validation check — so no depth figure is published. This public page explains and shows the physical signal; it does not expose the private interpretation built on top of it. Depth Fusion carries its own labelling — experimental, non-unique solution, not an ore body image — for the same reason this page carries the note above.
Gravity: Sandwell & Smith / TopEx free-air anomaly (Scripps Institution of Oceanography, UC San Diego). Magnetics: EMAG2v3 (NOAA NCEI). Scale and elevation-coupling figures are a GeaSpirit internal audit measured on the sources above; the method and raw samples are held in the GeaSpirit research record. This layer is geological context, not a resource or discovery claim.