How Satellite Spectral Analysis Actually Finds Copper Alteration Zones

Last March, I was staring at a Sentinel-2 scene over Chagai, Balochistan. A 12-kilometer patch lit up in band ratios I'd seen maybe twice before. Three weeks later, the field team confirmed phyllic alteration. Pyrite halos. The whole thing.

That's the part most people don't get about copper exploration from space. You're not looking for copper. You're looking for what copper did to the rock around it.

Copper doesn't sit alone — it cooks the neighborhood

When a porphyry copper system pushes hot fluids up through the crust, it changes everything it touches. The original rock — granite, andesite, whatever — gets chemically rearranged. Feldspars break down. New minerals form. We call these alteration zones, and they're usually bigger than the ore body itself. Sometimes 5 to 10 times bigger.

That's the gift. A copper deposit might be 800 meters across. The alteration halo around it? Could stretch 4 kilometers. And every one of those altered minerals has its own spectral fingerprint.

Here's the rough zoning you'll see around a classic porphyry (the Reko Diq style, basically):

• Potassic core — biotite, K-feldspar, magnetite. The hottest zone, closest to the intrusion.
• Phyllic zone — sericite, quartz, pyrite. This is the one satellites love.
• Argillic zone — kaolinite, illite, montmorillonite. Clay minerals, basically.
• Propylitic outer halo — chlorite, epidote, calcite. Cooler, wider, often overlooked.

Each of these absorbs and reflects sunlight differently. And that's where satellite spectral analysis earns its keep.

What the bands actually pick up

I'll keep this practical because honestly, most geology blogs make this sound harder than it is.

Sentinel-2 gives us 13 bands. The ones that matter for copper alteration are the SWIR bands — bands 11 and 12 — because clay minerals and sericite have strong absorption features around 2.2 micrometers. When we run a band 11/band 12 ratio, phyllic and argillic zones glow. Literally. Bright pixels where the alteration is strongest.

ASTER goes further in SWIR. Six bands between 1.6 and 2.43 micrometers. That's enough resolution to separate sericite from kaolinite from alunite, which Sentinel-2 can't really do on its own. For mapping argillic alteration specifically, ASTER's band 5 to band 6 ratio is something I run on almost every copper target we look at.

For iron oxides — the gossan caps that often sit above oxidized copper sulfides — we use the visible-near-infrared range. Band 4 over band 2 in Sentinel-2. Red over blue, basically. Hematite, jarosite, goethite. These show up bright orange and red in our processed scenes, and they're often the first thing a field geologist would spot from a helicopter anyway.

The trap I fell into early

When I started GeoMine AI, I thought spectral signatures alone would be enough. Run the ratios, find the anomalies, send the coordinates. Done.

It's not done. Not even close.

Clay minerals show up everywhere. Agricultural soils. Weathered shale. Riverbeds. Old construction sites. I had a scene over a part of Khuzdar that was screaming "argillic alteration" — turned out to be a clay quarry. Embarrassing.

So we layered in three things:

1. SRTM DEM data. Porphyry copper systems sit in specific structural settings. Faults, ring intrusions, topographic domes. If your spectral anomaly doesn't line up with a structural feature, it's probably noise.

2. SAR data. Radar tells us about surface roughness and moisture. Real alteration zones have distinct textures. Farm fields don't.

3. Multi-temporal stacking. We compare scenes across seasons. A real mineral signature stays put. Vegetation flux, moisture, agriculture — all of that changes month to month.

That's the geomines stack, basically. No single dataset is enough. But when Sentinel-2, ASTER, SRTM, and SAR all point at the same 2-square-kilometer patch? That's when I tell clients to send a team.

What this looks like in Pakistan

Pakistan has a copper problem and an opportunity at the same time. Reko Diq alone is sitting on an estimated 5.9 billion tonnes of ore. But the Chagai arc extends way beyond Reko Diq — there's roughly 480 kilometers of mineralized belt that's still under-explored. And the Waziristan ophiolite, plus parts of Kohistan, also host copper showings nobody's properly mapped.

The traditional way to explore this would cost a junior mining company maybe $2-3 million just for an initial campaign. Helicopters, camps, security, sampling crews. Most can't afford it. So the ground stays unexplored, and the country leaves money buried.

Satellite spectral analysis flips that math. We can pre-screen 10,000 square kilometers in a few weeks for a fraction of that cost. Generate ranked targets. Then the field budget goes to the top 5 or 6 anomalies instead of being sprayed across the map.

Look, I'm biased — I run a satellite intelligence company and I personally own 15 mines in Gilgit-Baltistan, so of course I think remote sensing matters. But the data backs me up. A 2019 study in Iran's Kerman copper belt showed ASTER-based phyllic mapping correctly predicted 78% of known deposits before any ground validation. That's not a small number.

A quick note on what spectral analysis can't do

It can't tell you grade. It can't tell you depth. It can't tell you tonnage. Anyone who says their satellite report does that is selling you something.

What it does is tell you where to drill first. Which in exploration economics, is most of the battle. Drilling a hole costs $80,000 to $250,000 depending on depth and access. Drill the wrong hole three times and your project is dead before it started.

So when somebody asks me what mineral remote sensing is really for, I tell them this: it's the cheapest decision-making tool you'll ever buy in mining. Not the answer. Just the question, asked sharper.

And in a country with $6 trillion under the ground and maybe 3% of it properly mapped — that sharper question is worth a lot.