Sentinel-2 vs Landsat: which to choose for your project in Colombia
Two free constellations, two different logics. We compare the resolution, frequency, bands and historical archive of Sentinel-2 and Landsat, and explain when each one fits —and why the answer is often to use both.
When a project needs free satellite imagery, it almost always comes down to two options: Sentinel-2 (European Space Agency) and Landsat (NASA/USGS). Both are optical, multispectral, openly accessible and globally covered. But they answer to different logics, and choosing wrong costs time: a historical series you can't build, a detail level that falls short, a frequency that leaves cloud gaps. This note compares the two in practical terms for the Colombian context.
The essentials in one table
| Feature | Sentinel-2 | Landsat 8/9 | |---|---|---| | Spatial resolution | 10 m (visible/NIR), 20 m, 60 m | 30 m (15 m panchromatic) | | Revisit frequency | ~5 days (two satellites) | ~16 days (8 days combining 8 and 9) | | Spectral bands | 13 (includes 3 red-edge) | 11 (includes thermal) | | Thermal band | No | Yes (TIRS) | | Historical archive | Since 2015 | Since 1972 (Landsat 1) | | Operator | ESA / Copernicus | NASA / USGS | | Cost | Free | Free |
Resolution: Sentinel-2 wins on detail
Sentinel-2 delivers 10 meters in the visible bands and near-infrared, versus Landsat's 30 meters. In practice, a Sentinel-2 pixel covers 100 m²; a Landsat pixel, 900 m². For delineating agricultural fields, tracking urban change or working at a coarse parcel scale, that gap is decisive. If your project is detail-driven, Sentinel-2 is the natural starting point.
Frequency: Sentinel-2 again, and it matters in the tropics
Sentinel-2 revisits the same spot every ~5 days; Landsat every 16 (every 8 if you combine Landsat 8 and 9). In Colombia, where cloud cover ruins most passes, frequency isn't a luxury: it determines how many usable images you get per year. More passes mean a higher chance of catching a clear day. For agricultural or environmental monitoring that needs temporal continuity, Sentinel-2 takes the lead again.
Where Landsat is irreplaceable
Landsat holds its own for two weighty reasons:
- Historical archive since the 1970s. If you need to analyze land-use change, deforestation or urban expansion across decades, Sentinel-2 (which started in 2015) simply doesn't have the data. Landsat is the only free source with half a century of continuity. For historical baseline studies, it's mandatory.
- Thermal band. Landsat measures surface temperature (TIRS), which Sentinel-2 doesn't. For urban heat islands, crop thermal stress, fire monitoring or water studies that depend on evapotranspiration, that band is key.
Bands: red-edge vs thermal
The most relevant spectral difference for agriculture: Sentinel-2 has three red-edge bands, ideal for indices like NDRE (nitrogen and stress in dense crops). Landsat has none. In exchange, Landsat brings the thermal band. So it's not that one has "more bands" and that's it: they have different bands, built for different questions.
The real answer: combine them
In serious projects, the question usually isn't "Sentinel or Landsat?" but "how do I combine them?" Two common strategies:
- Densify the time series. Using both constellations together multiplies available passes, critical for beating cloud cover. Harmonized products (like HLS, Harmonized Landsat-Sentinel) normalize both sources onto a common grid.
- Historical depth + recent detail. Landsat to reconstruct the past (decades-long baseline) and Sentinel-2 for fine monitoring of the present.
Quick guide by use case
| Need | Recommendation | |---|---| | Detailed agricultural monitoring | Sentinel-2 (red-edge, 10 m, 5 days) | | Historical land-use change (decades) | Landsat | | Surface temperature / heat islands | Landsat (thermal) | | Recent deforestation, high frequency | Sentinel-2 (+ Landsat to densify) | | Historical baseline + current tracking | Landsat + Sentinel-2 | | Field delineation / urban change | Sentinel-2 |
And when free isn't enough
Both Sentinel-2 and Landsat top out around 10 m. If the project demands meters or centimeters —physical cadastre, tree counting, fine infrastructure— you need to step up to high-resolution commercial imagery (PlanetScope, Pléiades, WorldView) or drone photogrammetry. The sensible rule: start with the right free source and step up only when the decision justifies it.
GeoSAT's role
Over 30 years we've worked with practically the whole ladder of optical sources, from Landsat and Sentinel-2 to commercial imagery and drone, choosing based on what each project actually needed. Selecting the right source —and the right combination— is one of the first technical decisions and the one with the biggest impact on cost and outcome.
If you're weighing which imagery to use for a monitoring, cadastre or territorial-analysis project in Colombia, get in touch: we define the question and the precision it demands first, and pick the source from there.