Gold Deposits in Gilgit Baltistan: What Satellite Geology Actually Suggests
I bought my first claim in Gilgit Baltistan in 2019. Back then I was reading geological survey reports from the 1970s and panning streams like everyone else. Now I run satellite scans across 47,000 square kilometers of the region in a single afternoon.
The difference is hard to overstate.
Let me tell you what the satellite data is actually showing about gold in GB — not the hype version, the version I'd tell another mine owner over chai in Skardu.
The Geology Nobody Outside Pakistan Understands
Gilgit Baltistan sits on one of the most violent tectonic collisions on earth. The Indian plate ramming into the Eurasian plate, with the Kohistan Island Arc squeezed between them like a sandwich. That collision zone — the Main Karakoram Thrust and the Main Mantle Thrust — is where the gold story begins.
Gold in GB shows up in two main settings. Placer deposits along the Indus, Gilgit, Hunza and Shyok rivers (which everyone knows about because locals have been panning them for centuries). And orogenic lode gold hosted in shear zones cutting through the Kohistan arc and the metasediments around Astore, Chilas and parts of Ghizer.
The placer stuff is obvious. The lode source — the actual mountain rock the placer gold eroded out of — is what satellite geology is finally helping us pin down.
Honestly, I used to think the placers were the prize. Took me three years and a lot of wasted helicopter fuel to realize the real money sits upstream, in the host rock, where nobody's drilling.
What Sentinel-2 and ASTER Are Picking Up
Gold itself doesn't show up on satellite imagery. Let's get that out of the way. No satellite on earth sees a gold atom from 700 km up.
But gold doesn't travel alone. It comes with friends — and those friends do show up.
At geomines we run multispectral analysis looking for the alteration mineralogy that wraps around orogenic gold systems. Sericite. Chlorite. Iron oxides like hematite and goethite. Silica flooding. Carbonate alteration where the host rock is mafic. ASTER's SWIR bands are particularly good at separating sericite from clay minerals — band ratios like (B5+B7)/B6 light up the phyllic alteration that often haloes a gold-bearing shear zone.
Three zones keep lighting up in our scans across GB:
- The corridor running from Bunji through Astore toward Rama — strong iron oxide anomalies coinciding with shear-zone lineaments visible in SAR data
- A belt north of Chilas where the Kohistan arc metavolcanics meet the Indus suture — classic orogenic gold setting, similar mineralogy to the Otago gold fields in New Zealand
- Patches in Ghizer district near Phander and along tributaries draining toward Gakuch, where hydrothermal alteration signatures are concentrated along fault intersections
SAR data — specifically Sentinel-1 interferometry — adds the structural piece. Gold loves structure. Fault intersections, splays off major thrusts, dilational jogs in shear zones. SAR sees those even when snow and vegetation hide them from the optical sensors.
SRTM DEM analysis then tells us which of those structures are actively eroding and feeding placers downstream. That's the validation loop. If a lineament in the highlands lines up with a placer-rich confluence below it, you've probably found the source.
What I Got Wrong At First
When we started building the geomine platform, I assumed the highest iron-oxide anomalies would be the best gold targets. Wrong.
Many of the brightest iron signatures in GB are just supergene weathering on ordinary metasediments — surface rust, basically. The real targets are the subtler ones where iron oxides sit next to sericite AND silica AND a mapped structural feature. Three-factor overlap. Not one bright pixel.
We also kept getting false positives over glacial moraines. The mineral mix in glacial debris mimics alteration signatures in a really annoying way. Took us about eight months of ground-truthing with samples from my own claims near Skardu and Astore before our models stopped chasing moraines.
That's the part of geo mining nobody writes papers about — the months of being wrong before the model starts being right.
The Numbers That Matter
GB's gold potential is real but I'm tired of the trillion-dollar headlines. Here's what I can defend with actual data:
- Recorded placer occurrences across the major river systems: 230+ documented sites
- Active artisanal gold extraction sites we've identified from high-resolution imagery: 1,847 (as of our last scan in October)
- Estimated lode-source target zones with three-factor satellite signatures: 73, of which maybe 12 are drill-ready in terms of access and structural confidence
The 12 number is the one I care about. Twelve targets in a region the size of Switzerland is not a gold rush. It's a careful, patient exploration program — the kind that built the Carlin Trend in Nevada over decades.
Why Nobody's Drilling Yet
Three reasons. Access is brutal — some targets sit at 4,200 meters with no road within 30 km. Licensing under the GB mineral rules takes patience and local relationships. And historically, exploration capital in Pakistan has chased copper and chromite because the unit economics are better understood.
Gold is changing that. At $2,600+ an ounce, the math on remote, high-grade orogenic systems suddenly works. A 5-gram-per-tonne shear zone that wasn't economic in 2015 is very economic now.
If you're sitting on a license in Astore, Diamer, Ghizer or Ghanche and you haven't run a proper satellite alteration study on it, you're flying blind. That's not a sales pitch — it's just true. The data is sitting there, free, in the Copernicus and USGS archives. Somebody just has to process it properly.
Which brings me to the question I keep asking the mine owners I meet up there: do you actually know what your license holds, or are you guessing based on what your grandfather found in the river?