Can satellites detect gold? What imagery can tell you

Three things to keep separate

Mineral information: some remote-sensing methods examine the way surface materials interact with measured radiation. They can support mineral mapping in appropriate conditions. That does not mean an ordinary image reveals a buried coin, container or gold deposit.

Indirect surface clues: vegetation, moisture and surface patterns can sometimes help researchers investigate buried historical features. Such clues require interpretation and may have several explanations. They are not a direct underground photograph.

Terrain information: an elevation model describes surface height according to the dataset's definition. Products differ in how they represent vegetation, buildings and the ground. Surface height is not the burial depth of an object.

These are explanations of general methods. They do not establish which sensors, datasets or analysis methods a particular application uses.

USGS: remote sensing of minerals ↗ · ESA: protecting history with satellites ↗ · USGS: elevation-model terminology ↗

Read the data before interpreting the picture

Start with the source and acquisition date, when available. Identify whether you are looking at a photograph-like image, an elevation layer, a shaded rendering or a calculated visualization. A realistic-looking 3D view can still be a representation of data rather than an observation of the underground.

Check the scale and resolution. A raster elevation model records values on a grid; enlarging its display does not create measurements between the original samples. A feature smaller than a cell cannot be explicitly represented as its own resolved feature in that raster. Keep unknown resolution marked as unknown.

Keep acquisition date separate from the date you opened the map. A recently viewed image is not necessarily a recent observation. If a comparison uses different dates, record that difference before treating a change as meaningful.

USGS: elevation-model resolution ↗

A practical way to evaluate an interesting pattern

1. Describe only what you can see. Write “a darker rectangular patch beside a field boundary”, rather than “a buried chamber”.
2. Record the context. Note the source, date, approximate extent and whether the feature follows a road, slope, drainage line or land-use boundary.
3. List alternatives. A difference might reflect vegetation, moisture, shadow, land use, geology or the way the data were processed. Do not choose an explanation merely because it is exciting.
4. Look for a useful independent check. Another observation should test a specific explanation. Repeating the same visualization with different colors does not supply a new measurement.
5. Keep the conclusion proportional to the evidence. “A feature worth investigating” is a valid outcome. “Unresolved” is also a valid outcome.

Worked example: a rectangular patch

This is an imaginary teaching example, not a SHAMAL discovery record.

You notice a dark rectangular area in an image. The observation is its shape and contrast. Possible explanations include a land-use boundary, differing vegetation or moisture, or another surface condition. A hidden structure is a hypothesis, not a fact established by the shape.

You would first record the image date and scale, inspect surrounding boundaries and ask whether the pattern persists in suitable independent observations. If those checks remain inconclusive, keep the interpretation open. Do not assign a material, depth or success percentage from the outline.

Where a metal detector or another field method fits

A field instrument and an image interpretation provide different observations. A useful field investigation begins with a specific question and a method appropriate to that question. Keep its settings, conditions, location reference and result separate from the image record. Agreement between two observations can be informative, but it does not automatically identify gold or establish a discovery.

Before any visit, check access, permission and applicable local requirements. An interesting image provides no right to enter land or disturb a site. Use suitable professional advice for questions that require specialist investigation.

A better record than a confident label

For each observation, retain:

The aim is a record another person can understand and assess. A color, model or confident caption cannot replace that evidence.

• The source, acquisition date and relevant credit.
• The data type, scale and resolution, or an explicit “unknown”.
• A plain description of the visible feature.
• Your interpretation and at least one plausible alternative.
• Independent evidence already available and what it actually shows.
• The next question to test and the current uncertainty.