EMAG2 Global Magnetic Data: The Free Resource Most Explorationists in Pakistan Are Sleeping On

By Sufyan · 2026-04-29 · 4 min read

Last month I sat with a mine owner from Chagai who'd just paid 4.2 million rupees for a magnetic survey over a 12 sq km block. Decent survey. Useful data. But when I pulled up EMAG2 on my laptop and showed him the regional magnetic signature of his entire district — for free, in about ninety seconds — his face dropped.

He'd been exploring blind for three years. Nobody told him this dataset existed.

So let's fix that.

What EMAG2 Actually Is

EMAG2 stands for Earth Magnetic Anomaly Grid, 2 arc-minute resolution. It's a global compilation built by NOAA, GFZ Potsdam, and a handful of other agencies. They stitched together satellite measurements, marine surveys, and airborne magnetic data from across the planet into one continuous grid. Version 3 came out in 2017 and it's still the best free magnetic dataset covering Pakistan end-to-end.

Resolution is 2 arc-minutes. That's roughly 3.7 km per pixel at our latitude. Not enough to drill on. More than enough to decide where to fly a detailed survey, or where to send a field team, or which 50 sq km block out of a 5,000 sq km license area deserves your attention first.

And it costs zero rupees. You download it from NOAA's NCEI portal. That's it.

Honestly, when I first started GeoMine AI I underestimated this dataset. I was chasing Sentinel-2 spectral signatures and ASTER SWIR bands and barely glanced at magnetics. Then we ran a back-test on a known chromite belt in Muslim Bagh and the EMAG2 anomaly lined up almost perfectly with the producing zones. I'd been ignoring a free regional filter that could've saved us months on three different projects.

Why Magnetic Data Matters for the Minerals We Care About

Magnetic surveys work because different rocks have different magnetic susceptibilities. Magnetite-rich rocks light up. Quartz and limestone barely register. Faults and shear zones often show up as linear discontinuities in the magnetic field — and faults are where hydrothermal fluids travel, which is where gold and copper deposits form.

For Pakistan specifically, here's what global magnetic anomaly data tells you fast:

Lithium pegmatites? Magnetics is less direct there — pegmatites are usually magnetic lows in a granitic background. But combined with our Sentinel-2 lithium workflow, EMAG2 still narrows the search area meaningfully.

How We Actually Use It at GeoMine AI

Our pipeline isn't complicated. It's just layered.

Step one: pull the EMAG2 grid for the client's area of interest. We clip it, reproject to UTM Zone 42N or 43N depending on where in Pakistan we are, and run a reduction-to-pole correction because Pakistan sits at a magnetic inclination where raw anomalies get shifted off the source bodies. Skip this step and your interpretation is off by a few kilometers in the wrong direction. I got this wrong on an early Balochistan project and it took me two weeks to figure out why our magnetic highs weren't matching the surface geology.

Step two: derivative filters. First vertical derivative sharpens edges. Tilt derivative highlights structural fabric. Analytic signal gives you body locations regardless of magnetization direction. These are standard processing steps in any geomining workflow but most people in Pakistan aren't doing them on free datasets — they're paying consultants to do it on expensive datasets.

Step three: stack EMAG2 against everything else. SRTM DEM for topography. Sentinel-2 for surface alteration. ASTER for mineral mapping. SAR for structural lineaments through cloud cover. When four independent datasets all point at the same 2 sq km zone, that's where you spend money.

Step four — and this is where the AI comes in — we train models on known deposit signatures across the region and let them flag look-alike patterns elsewhere. EMAG2 is one input feature among many, but it's a feature we can get for every square kilometer of Pakistan without negotiating with anyone.

The Catch

Look, EMAG2 isn't magic. The 3.7 km resolution means you'll miss small high-grade deposits that don't have a regional footprint. A narrow gold-bearing quartz vein in Kohistan won't show up. A 2 km wide alteration halo around a porphyry system probably will.

It also has data quality variations. Some parts of Pakistan were surveyed properly decades ago. Other parts are interpolated from sparser measurements and the noise is real. The Indus plain is well-covered. Parts of upper Chitral and the Karakoram interior — less so.

But for a country where most exploration licenses are awarded over areas larger than 100 sq km, and where most license holders have no idea where to start within their own block, free magnetic data mining is genuinely the cheapest first move you can make. The download is 200 MB. The processing takes an afternoon if you know what you're doing.

The Geological Survey of Pakistan has better data than EMAG2 for some districts. Getting access is its own adventure. In the meantime, the global dataset sits there, public, ignored.

If you're holding a mining license right now and you haven't looked at the EMAG2 signature of your block, you're making decisions with one eye closed. Why would you do that when the other eye is free to open?

© 2026 GeoMine AI. All rights reserved.