By Laura E.P. Rocchio, Landsat Communication and Public Engagement Team
For more than a half-century, Landsat satellites have recorded data about Earth’s land surface, giving scientists the ability to monitor changes across time and landscapes. With each generation of Landsat sensors, the instrumentation has become more advanced and their measurements more precise.
Careful attention to Landsat data calibration, algorithm development, and end-user products have led to widespread use of the data for global time-series analysis, uncovering global trends in many biophysical domains—fire, agriculture, water use, forest cover, urban expansion, and glacier recession.
Renowned atmospheric physicists have developed atmospheric corrections for Landsat that account for the intervening atmosphere between the satellite and Earth surface. These corrections are implemented to create robust surface reflectance products used for global trending, the most recent of which is the USGS Landsat Collection 2 Level 2 Surface Reflectance product.
As its name suggests, Landsat was developed to monitor the Earth’s land surface, but the improved radiometric resolution and spectral coverage of Landsats 8 and 9 have provided a new avenue for water-column science targeting water transparency, in-water biomass, organic matter, suspended sediment, and the like. While floating algae, emergent aquatic vegetation, and historic surface scum can be tracked throughout the Landsat record, older Landsat sensors lack the precision needed to be used for water-column studies.
A recent letter in Limnology and Oceanography details that caution must be applied when using the Collection 2 Surface Reflectance products for water-column science as this domain requires an aquatic-specific, or “water-tuned” atmospheric correction. Users are advised to instead use the provisional Aquatic Reflectance product available from USGS for water-column studies. They are also advised not to use the Surface Reflectance products to conduct long-term water-column trend analysis with satellite data prior to Landsat 8, unless they use band ratios or other methods to adjust for observed biases.
To reach these conclusions, the study authors used a subset of ~12,000 in situ spectral reflectance measurements (largely from the open-access GLORIA dataset) to evaluate the quality of Collection 2 Surface Reflectance for water-column studies. When comparing in situ data collected within one day of Landsat overpasses across numerous geographic locations, authors found that Surface Reflectance products for Landsats 5 and 7 lack the precision to be used in rigorous water-quality studies—overestimating water body reflectance by more than 25 percent. This overestimation means that false downward trends will be found when using Landsat 5 and 7 Surface Reflectance time-series data in combination with that from Landsats 8 and 9.
Landsat’s human-scale spatial resolution makes it an important resource for inland water studies, and the next-generation satellite, Landsat Next, will be even better suited for water-column studies than Landsats 8 and 9. Study co-author and a 2018-2023 Landsat Science Team member, Nima Pahlevan, shares that Landsat Next’s new 412 nm spectral band will measure specific aerosol contributions that impact water body reflectance, making better water-tuned atmospheric corrections possible.
USGS and NASA are committed to continued algorithm development and validation for a consistent and credible Landsat record for water-column relevant studies, and plan to support new atmospheric correction research for Collection 3 and Landsat Next.
“The Landsat mission is only at the genesis of provisioning, operational systematic high quality aquatic remote sensing data products,” shared Chris Crawford, USGS Landsat Project Scientist. “Robust atmospheric correction techniques coupled with quantified observational uncertainties, is central to enabling broad use of the Landsat archive for long term assessment and monitoring of Earth’s coastal and inland water bodies.”
Reference
Maciel, D. A.; Pahlevan, N.; Barbosa, C. C. F.; de Moraes de Novo, E. M. L.; Paulino, R. S.; Martins, V. S.; Vermote, E.; Crawford, C. J. Validity of the Landsat Surface Reflectance Archive for Aquatic Science: Implications for Cloud-Based Analysis. Limnology and Oceanography Letters 2023, https://doi.org/10.1002/lol2.10344Related Reading
Behind the Paper: GLORIA – Challenges in developing a globally representative hyperspectral in situ dataset for the remote sensing of water resources; Earth & Environmental Sciences Community