A reference measurement is an independent measurement of a quantity, traceable to a recognised standard where possible, that is used as the basis for calibrating or validating a satellite observation.12

The CEOS quality framework asks that quality indicators be traceable to reference standards, preferably SI, and links this to uncertainty analysis and comparison. The CEOS calibration and validation group offers the community access to reference data, networks and internationally agreed protocols. None of the sources opened gives a single formal definition, so this description is drawn from how they use the term.12

For validation, a reference is usually collected on the ground and then related to the satellite footprint. A National Physical Laboratory report states that the sampling scheme must give an aggregated value that does not depend on the location of the samples, and it focuses on optical products over tropical forests.3

A reference is not automatically adequate. NASA presents Landsat calibration as independently verified against ground conditions, but whether a given reference suits a given product depends on its uncertainty, representativeness and timing, which the sources do not settle in general.1234

Footnotes

  1. CEOS QA4EO, “QA4EO Home”. https://qa4eo.org/ ↩ ↩2 ↩3

  2. CEOS, “Working Group on Calibration and Validation (WGCV)”. https://ceos.org/ourwork/workinggroups/wgcv/ ↩ ↩2 ↩3

  3. National Physical Laboratory, Sampling and scaling for satellite biophysical product validation from optical sensors with an emphasis on tropical forests, NPL Report OP 11, 2015. https://eprintspublications.npl.co.uk/6571/1/OP11.pdf ↩ ↩2

  4. NASA Science, “Landsat Calibration and Validation”. https://landsat.gsfc.nasa.gov/data/calibration/ ↩