SRTM DEM Explained: How Elevation Data Predicts Marble and Granite Quarries in Pakistan

Last March, one of my field guys in Chillas sent me a WhatsApp voice note. He'd walked 11 kilometers up a ridge my team had flagged from elevation data. Found a marble face the size of a cricket ground. Nobody had touched it.

That's what SRTM DEM does when you actually know how to read it.

Most people in the Pakistani mining scene still treat elevation data like it's a Google Earth toy. Zoom in, look at the mountain, guess where the rock is. But SRTM (Shuttle Radar Topography Mission) data — collected by NASA back in 2000 using radar interferometry from the Space Shuttle Endeavour — gives you elevation values at 30-meter resolution for almost every corner of Pakistan. Free. And when you combine that with what we already know about how marble and granite form, you can predict quarry sites before anyone climbs a single ridge.

Let me explain how.

Why Elevation Actually Matters for Marble and Granite

Marble is metamorphosed limestone. It sits where old sedimentary rock got cooked and squeezed — usually along fold belts, thrust zones, and ridges that were pushed up during tectonic collisions. In Pakistan, that means the Indus Suture Zone, the Kohistan arc, and the fold belts running through Chitral, Buner, and parts of Mohmand.

Granite is different. It's igneous, cooled from magma deep underground, and it shows up as batholiths — huge dome-shaped intrusions. The Mansehra granite, the Karakoram batholith around Hunza, and the Nagar Parkar granites in Sindh are the big three most people know.

Here's the thing both rock types have in common: they resist erosion. Softer rock around them washes away over millions of years. What's left behind sits higher, steeper, and shows up in elevation data as distinct patterns.

Marble ridges typically show as long, narrow crests — sometimes only 40 to 90 meters wide but running for kilometers. Granite bodies show as broad, rounded domes with slope angles between 22 and 38 degrees on their flanks. Once you've looked at enough SRTM data, you start seeing them without even trying.

What We Actually Do With the Numbers

SRTM DEM mining analysis isn't magic. It's math applied to elevation.

We pull the raw DEM tile for a target area — say, a 50 km by 50 km block in Mohmand district where marble is already known. Then we run four things:

Slope analysis. Marble outcrops usually sit on slopes between 25 and 45 degrees. Too flat and it's alluvial cover. Too steep and it's probably shale or phyllite that broke apart. We filter the DEM to show only pixels in that slope range.

Aspect analysis. This one surprised me when I first got into it. Marble faces exposed to certain directions weather differently. In northern Pakistan, south and southwest-facing slopes tend to show cleaner exposures because of freeze-thaw patterns.

Curvature. Convex ridges (positive curvature) suggest resistant rock underneath. Concave valleys eat away softer material. Marble ridges almost always show up as strong positive curvature lines.

Drainage pattern extraction. Granite bodies create radial drainage — streams flowing outward like spokes on a wheel. Metamorphic terrain with marble creates trellis or rectangular drainage. Just by looking at the stream network we derive from SRTM, we can often tell what rock type dominates.

At geomines, we combine all four layers with Sentinel-2 reflectance and ASTER thermal bands. But honestly, even before we add the multispectral data, the SRTM analysis alone gives us 60-70% confidence on where to look. That's a lot better than sending a jeep up a random valley and hoping.

A Real Example From Buner

Last year we ran a marble quarry Pakistan prediction study across Buner district. The client had two operational quarries and wanted to know where to buy adjacent land before prices moved.

We pulled the SRTM tile, ran the slope-curvature-drainage stack, and cross-referenced with the known marble outcrops. The signature was clear: elongated ridges at 1,400-1,900 meters elevation, slope 28-40 degrees, strong positive curvature, trellis drainage.

Then we let the model find every other pixel cluster in Buner matching that signature. It came back with 17 candidate zones. We ground-truthed 6 of them with a local geologist over three weeks. Four had exposed marble. One had marble under about 2 meters of soil cover. One was a bust — turned out to be a quartzite ridge that looked identical from space.

That's a 5-out-of-6 hit rate on ground verification. From a laptop in Islamabad.

I won't pretend I got this right the first time I tried it. Back in 2022 when I was still figuring out our geomine workflow, I ignored curvature completely and just used slope. My hit rate was maybe 30%. Adding curvature and drainage extraction — which took me weeks of reading papers from the Pakistan Geological Survey and playing with QGIS — is what pushed accuracy into a useful range.

Where SRTM Falls Short (Because It Does)

SRTM is 30-meter resolution. That means anything smaller than a 30x30 meter feature just disappears. Small marble lenses, narrow granite dikes, tight veins — you won't see them.

Also, SRTM data was collected in February 2000. Twenty-five years of erosion, road building, and quarrying since then means the elevation values are slightly off in mined areas. Not a huge issue for exploration, but worth knowing.

And SRTM can't tell you rock chemistry. It shows you where resistant rock probably sits. Whether that rock is white Ziarat marble worth $180 per ton at the quarry gate, or a dirty grey stone worth almost nothing, requires spectral analysis. That's where our breeze geo mineral analysis workflow at geomines picks up — combining the DEM prediction with hyperspectral signatures to sort the good rock from the junk.

Look, if you own land in Mohmand, Buner, Chitral, Mansehra, Hunza, or anywhere along the northern fold belts, running SRTM analysis on your property before you spend a rupee on drilling is just common sense. The data is free. The processing takes a couple of days. And the alternative — sending crews up mountains on hunches — is how mining companies in Pakistan have been burning money for the last forty years.

What's the elevation of your ridge?