Landsat for Mineral Exploration: Still Relevant in the Sentinel Era?

Someone asked me last week why we still pull Landsat scenes at GeoMine AI when Sentinel-2 has better resolution, faster revisit times, and sharper bands for mineral mapping.

Fair question. And honestly, I used to think the same thing — that Landsat was the old workhorse we'd eventually retire. Then I spent 18 months running comparisons across the Chagai belt and parts of Gilgit-Baltistan. I changed my mind.

Let me explain why.

The Thing Sentinel-2 Can't Give You

Landsat has been shooting pictures of Earth since 1972. That's 52 years of continuous imagery. Sentinel-2 launched in 2015.

Think about what that means for a copper prospect in Balochistan. If I want to see how vegetation patterns, iron oxide signatures, or surface alteration have shifted over four decades — whether mining activity is expanding, whether rivers have cut new exposures, whether a landslide revealed fresh rock — Sentinel-2 gives me 9 years of data. Landsat gives me half a century.

For mineral exploration, time depth matters more than people admit. A single cloudy year in Sentinel can leave a 12-month gap. Landsat's archive fills it.

And here's the thing — the USGS made the entire Landsat archive free in 2008. Every single scene. That decision quietly changed exploration geology forever, and most people outside the field still don't know it happened.

Landsat vs Sentinel: What Actually Matters on the Ground

Let's get specific because the generic comparisons online are mostly useless.

Sentinel-2 has 13 bands at 10-20 meter resolution, 5-day revisit. Landsat 8 and 9 have 11 bands at 30 meter resolution (15m panchromatic), 16-day revisit combined.

On paper, Sentinel wins. Higher resolution. Faster revisit. More red-edge bands that are genuinely useful for vegetation-based indirect indicators.

But Landsat has something Sentinel doesn't — thermal infrared bands. TIRS-1 and TIRS-2. These read surface temperature, and for certain mineral mapping work, especially around hydrothermal systems and silica-rich alterations, thermal data is not optional. It's the whole point.

When we were profiling a chromite zone near Muslim Bagh last year, the Sentinel-2 data showed us iron oxide signatures clearly. But the thermal anomalies picked up by Landsat 9 — a 3.2°C differential across a 4km stretch — pointed to a structural feature that Sentinel simply couldn't see. We confirmed it on the ground six weeks later. Sheared ultramafic contact. Classic chromite host structure.

Sentinel didn't miss it because it's bad. It missed it because it doesn't carry the sensor.

So when someone asks me Landsat vs Sentinel for Landsat mineral exploration work, my answer is boring — use both. They're not competitors. They're different instruments answering different questions.

Where Landsat Remote Sensing Mining Still Wins

A few scenarios where I reach for Landsat first, not Sentinel:

Historical baseline work. If a client wants to know what a license block looked like in 1995 before illegal mining started, Sentinel can't help. Landsat 5 was up there watching.

Thermal mapping for geothermal-adjacent mineral systems. Pakistan has under-explored geothermal zones in Chitral and the northern areas that correlate with specific mineralization patterns. Thermal bands are how you find them from space.

Large regional surveys where 30m resolution is enough. If I'm scanning a 50,000 sq km area in the first pass, I don't need 10m pixels. I need consistency, wide swath coverage, and an archive I can run time-series analysis against. Landsat's 185km swath and 40-year record beat Sentinel for that specific job.

Atmospheric correction stability. The Landsat calibration pipeline is older, more tested, and more trusted by geologists who've been doing this work since before half my engineers were born. That matters when you're handing a report to a mining executive who needs to sign a $40 million investment decision.

Look, I'm not romantic about satellites. I'm romantic about getting the answer right. At geomines we built our stack around the idea that no single sensor tells the whole story. Sentinel-2 for detail. Landsat for depth and thermal. ASTER for SWIR mineral mapping (and yes, ASTER's archive is basically frozen since 2008 for SWIR, which is a separate headache). SAR for structure and through-cloud imaging. DEM for terrain and drainage.

Any exploration company telling you they only need one data source is either selling something or they haven't been burned yet.

The Honest Verdict

Is Landsat still relevant? Yes, but not for the reasons people usually give.

It's not relevant because of resolution — Sentinel clearly wins there. It's not relevant because of revisit time — Sentinel wins that too. It's relevant because of three specific things: thermal infrared bands, a 52-year archive, and radiometric consistency that lets you do serious time-series analysis without fighting calibration drift.

For a mining company operating in Pakistan — whether you're prospecting copper in Reko Diq territory, chasing lithium pegmatites near Mansehra, or trying to verify what's actually on your Gilgit-Baltistan lease — ignoring Landsat because it's older is like ignoring a geological report because it's written in pen.

The data doesn't age. Only our attention spans do.

What I'd actually tell a mine owner asking where to start? Pull both. Run them together. Let the overlaps confirm and the disagreements raise questions. That's where the real prospects hide — in the places where two sensors tell you slightly different stories and you have to figure out which one is lying.

Usually neither is. You're just not asking the right question yet.