Landsat 5/7/8/9 Cross‑Sensor Harmonization

Last updated: May 5, 2025

Applies to: Landsat Collection 2 Level‑2 Tier 1 Surface Reflectance (SR) and Collection 2 Tier‑1 Top‑of‑Atmosphere (TOA) products in the Climate Engine App and API

Summary

Climate Engine does not apply additional cross‑sensor harmonization when you request spectral or thermal bands, NDVI, EVI, or other band‑based metrics from Landsat 5, 7, 8, or 9 beyond the corrections applied to Landsat Collection 2.
The USGS Collection 2 processing pipeline already standardizes radiometry, geometry, and per‑scene quality so that comparisons across sensors are considered analysis-ready for most applications.
Published guidance from Google Earth Engine (GEE) and the USGS Landsat team indicates that standardized harmonization coefficients have not been produced for Collection 2 and are likely unnecessary.

Why doesn’t Climate Engine harmonize Landsat sensors?

Collection 2 is generally considered analysis-ready.

The USGS reprocessed the entire Landsat archive using similar atmospheric‑correction model s(LaSRC for optical SR for OLI and LEDAPS for TM and ETM+).
From USGS: "Landsat scenes with the highest available data quality are placed into Tier 1 and are considered suitable for time-series analysis. Tier 1 includes Level-1 Precision and Terrain (L1TP) corrected data that have well-characterized radiometry and are inter-calibrated across the different Landsat instruments. The georegistration of Tier 1 scenes is consistent and within prescribed image-to-image tolerances of ≦ 12-meter radial root mean square error (RMSE)."
Level‑2 SR products are delivered in the same spatial grid (30 m), using the same metadata structure and Quality Assessment (QA) bits.
From USGS: "Landsat Collection 2 Level-2 Science Products are time-series observational data processed for consistency and continuity to measure effects of environmental change and serve as input into Landsat essential climate variable Level-3 Science Products."

Guidance and lack of evidence for the need for additional harmonization.

Earth Engine’s FAQ on cross‑sensor SR states that additional harmonization is usually unnecessary.
From the above article: "Roy et al., 2016 included an analysis of reflectance differences between Landsat 7-8 TOA and surface reflectance.... The improvements made during Collection 1 and Collection 2 reprocessing may influence the relationship between sensors, but as far as we know, there have been no analyses similar to Roy et al. (2016) for Collection 1 or Collection 2 data"

Practical experience in using Landsat Collection 2

In our experience, you may notice small differences in surface reflectance between Landsat 5 and 7 for overlapping time periods (1999-2001), and between Landsat 5 (ending 2012) and Landsat 8-9 time periods (2013-present) but these differences are very small and only noticeable in very sparse vegetation (e.g. desert areas where NDVI < 0.1).

Frequently‑asked Questions

Q1. Does Collection 2 harmonize TOA products too?

Yes. Collection 2 Tier‑1 TOA reflectance and radiance are radiometrically recalibrated using updated gain/bias tables. However, TOA values are still influenced by atmosphere; we recommend SR for biophysical indices.

Q2. Is harmonization different for across indices?

No. All indices are computed from the same spectral bands. Any residual cross‑sensor bias in band reflectance will propagate similarly into any index.

Q3. What about Landsat 1‑4 MSS or Collection 1 data?

Older missions and previous collections were not processed with the same algorithms. Cross‑sensor calibration is more complicated and not covered by this FAQ. We do not provide MSS data in Climate Engine at this time and Collection 1 has been superseded by Collection 2.