akar.earthLandsat: Definition
What is Landsat? The longest-running Earth observation satellite programme, and how its archive supports forest baselines and land history checks in Indonesia.
Landsat is a series of Earth observation satellites run by NASA and the US Geological Survey that has imaged the planet’s land surface continuously since 1972. Its free archive of 30 m multispectral imagery is the main record of how forests have changed over the past five decades.
Why it matters
Forest carbon accounting needs history. REDD+ baselines rely on past deforestation rates, ARR projects must prove land was not cleared in the years before the project, and many global forest change datasets are built on Landsat. No other free source covers such a long period at a useful resolution.
How it is used
- Historical deforestation and degradation analysis for baseline setting.
- Land eligibility checks, such as showing a site was non-forest for the required period under the 10-year rule.
- Long time series of vegetation indices to show recovery or decline.
- Harmonised with Sentinel-2 to give denser recent time series.
Recent Landsat satellites carry improved sensors and, with two in orbit, revisit any point more often than the 16 days of a single satellite.
Indonesia and Southeast Asia angle
Much of what is known about Indonesia’s forest loss since the 1990s, including the expansion of oil palm and pulpwood and the impact of the 1997–98 and 2015 fires, comes from Landsat analysis. For projects, the archive helps document land use history on degraded land, former concessions and peat. Cloud cover is a problem here too, and older scenes are sparser, so some years may have gaps. See drivers of deforestation in Indonesia.
Because Landsat sensors have changed over time, long time series need harmonisation so that a change in pixel values reflects the land, not the instrument. Analysis-ready collections from USGS handle much of this.
Common mistakes
- Expecting 30 m pixels to show small smallholder plots or individual trees.
- Ignoring the scan-line gaps in Landsat 7 imagery after 2003.
- Mixing sensors without radiometric harmonisation.
Landsat is one tool in the wider remote sensing toolkit.
Frequently asked questions
What resolution is Landsat?
Most Landsat multispectral bands are 30 m, with a 15 m panchromatic band on recent satellites. That is coarser than Sentinel-2 but covers a much longer history.
Why use Landsat if Sentinel-2 is sharper?
Landsat goes back to the 1970s and 1980s. For baselines, historical deforestation rates and land eligibility checks, you often need imagery from before Sentinel-2 launched in 2015.